This study identified characteristics of patients with colorectal cancer who traveled farther for surgery and found that those who traveled tended to stay longer at the hospital.
Objectives: This study sought to examine the impact of distance traveled from place of residence to surgical facility for elective colorectal surgery on surgical outcomes, length of stay, and complication rate.
Study Design: Retrospective study.
Methods: Patients with colorectal cancer were identified from the Florida Inpatient Discharge Database. Distance traveled from primary residence to surgical facility was estimated using zip code. After adjusting for patient and hospital characteristics, multivariate regression models compared bypassed hospitals, the length of stay, and complication rates for patients traveling different distances to receive care.
Results: Patients residing in rural areas and in South (odds ratio [OR], 2.37; 95% CI, 1.55-3.63) and Central Florida (OR, 5.86; 95% CI, 3.86-8.89) were more likely to travel more than 50 miles for treatment. Teaching status of the hospital (OR, 9.99; 95% CI, 6.98-14.31), a hospital’s availability of a colorectal surgeon (OR, 1.83; 95% CI, 1.45-2.31), and metastasized cancer (OR, 1.43; 95% CI, 1.17-1.82) influenced the patient’s decision to travel farther for treatment. Length of stay was significantly higher for patients traveling farther (P < .0343). However, there was no significant difference in the rate of complications among the groups (those traveling 25-50 miles vs < 25 miles [P = .5766] and those traveling > 50 miles vs < 25 miles [P = .4516]).
Conclusions: A greater number of patients travel more than 50 miles to the surgical facility at a later stage of disease. These patients do not significantly differ from those traveling less than 50 miles in their rates of complications; however, they stay longer at the surgical facility.
Am J Manag Care. 2020;26(11):e347-e354. https://doi.org/10.37765/ajmc.2020.88529
This study identified characteristics of patients with colorectal cancer who traveled farther for surgery and found that those who traveled tended to stay longer at the hospital.
For patients who face a diagnosis of cancer, the ability to access needed care can influence health care outcomes and survival.1 Because optimal care is often found at specialty regional centers or may be offered at locations that are not convenient to patients’ homes, there is a wide differential of travel burden experienced by patients. Although patients expect to receive the same quality of care, variances in measures of health care access and outcomes are often observed for those traveling greater distances for treatment.2,3 In addition, disparities in quality measurements have been demonstrated by studies in a variety of care settings.1,2,4 For example, in a study utilizing the National Cancer Database, patients with gastrointestinal cancers who traveled farthest for intent-to-cure surgery—regardless of facility or demographic or socioeconomic status—had lower risk of 90-day and 5-year mortality.5
Length of stay (LOS) at surgical facility and in-hospital complications are major drivers of health care cost for patients and are commonly used metrics in comparative effectiveness research.6,7 Although the impacts of clinical factors on quality metrics of perioperative care have been widely studied in patients with colorectal cancer,8 less attention has been dedicated to identifying impacts of nonclinical factors, such as travel and access to care. Identification of the relationship between accessibility of care and LOS and in-hospital complications is crucial because it provides additional insights for proper allocation of resources to bolster services wherever needed. This study was conducted to identify the factors that cause patients with colorectal cancer to travel farther for surgical treatment and to further explore the impact of these factors on LOS and in-hospital complications.
This study utilizes data from the 2013-2015 Florida Inpatient Discharge Database. This data set includes deidentified patient-level data from all acute care hospitals within the state of Florida, including information regarding cancer diagnosis, metastatic spread, treatment, and hospital and physician identifiers.9 This database does not include protected health information and hence the study was exempted from a formal review by the Mayo Clinic Institutional Review Board. All patients receiving inpatient colorectal cancer surgery during the years 2013 to 2015 were included in the evaluation. Patients were excluded if they were admitted urgently or emergently, because these patients are likely to seek care from the nearest surgical facility.
The dependent variables for this study were hospital chosen vs bypassed, LOS, and surgical complications. For patients who traveled greater than 50 miles (80.5 km) for care, a binary variable was created identifying the hospital in which they received care compared with the hospitals in the 25-mile (40.2-km) radius of the patient’s home zip code that they bypassed. LOS was operationalized as a continuous variable. Surgical complications were operationalized as a binary variable based upon the Healthcare Cost and Utilization Project Clinical Classification Software (CCS) 238 codes, which identify medical and surgical complications. In addition, CCS codes for thromboembolic complications including deep venous thrombosis and pulmonary embolism (415, 415.11-415.19, 4510-4519, 4530-4539) were used to identify postoperative complications.10,11 If a patient experienced any of the complications associated with the surgical procedure of interest, they were classified as having a surgical complication.
Distance traveled from the primary residence to surgical facility was estimated by calculating the distance between the zip code centroids for the patient’s residence and the surgical facility. We created a 3-category exposure variable—travel distance—with categories of (1) less than 25 miles, (2) between 25 miles and 50 miles, and (3) more than 50 miles. The cutoff points for the travel distance were selected based on the knowledge that the median (interquartile range) travel time of the US population to the nearest academic cancer center is 30 (13-72) minutes.12 Travel time of more than an hour, corresponding to approximately 50 miles (80.5 km) to the surgical facility, was classified as “distant travel,” and travel time of less than 30 minutes, corresponding to 25 miles (40.2 km), was classified as “local travel.” The third category included patients who traveled a moderate distance, between 25 miles and 50 miles.
Patient covariates included age, sex, race/ethnicity, rural/urban location, residential region in Florida, payer type, and comorbidities. Hospital factors included hospital size, hospital volume, teaching status, and availability of a colorectal surgeon. Race/ethnicity were categorized as White, Black or African American, and Hispanic or Latino. Payer type included commercial, Medicaid, Medicare, Medicare Managed Care, and other, which comprised self-payers and uninsured. The comorbidity burden was measured using the Elixhauser score, and 4 categories were created indicating the presence of 0, 1 or 2, 3 to 5, or more than 5 comorbidities.13 Metastatic status of the cancer and obesity were included separately due to their important influence on a patient’s ability to travel. Hospital size was identified using the number of hospital beds: Fewer than 100 indicated a small hospital; 100-299, a medium hospital; and 300 or greater, a large hospital. Hospital volume was analyzed both as a categorical variable, based on the quartile distributions of the number of colorectal surgical procedures performed at each hospital in the data set (0-48, 49-105, 106-149, and 150-277), and as a continuous variable. Teaching status of the hospital was identified based on the presence of an Accreditation Council for Graduate Medical Education residency program.
Bivariate analysis examined associations among patient characteristics, hospital factors, and the 3-category variable travel distance. Pearson’s χ2 test and the Kruskal-Wallis test were used to compare categorical and continuous variables, respectively. Pairwise comparisons were conducted to understand characteristics of patients traveling farther for treatment. Multivariate logistic regression was used to examine hospital-level factors associated with chosen facilities vs those bypassed. Multivariate linear regression was performed to compare the LOS at surgical facilities for patients who traveled farther for surgery with that of those treated locally, after adjusting for other covariates. Due to the nonnormal distribution of the LOS variable, logarithmic transformation was utilized to meet regression assumptions. Multivariate logistic regression was performed to understand the complication rate for patients by the travel groups. Finally, a sensitivity analysis was conducted to examine whether any of the selected covariates was an effect modifier of the association between travel distance and LOS or complication rate.
A total of 8798 patients with colorectal cancer were included in the analysis (Table 1 [part A and part B]). Of the total, 5.95% of patients traveled more than 50 miles; 8.2%, between 25 and 50 miles; and 85.8%, less than 25 miles to receive surgery. The mean age was 68.4 years. The majority of patients were White (75.0%), followed by Hispanic (15.4%) and African American (9.6%). Patients from rural areas made up 5.4% of our study sample, and 19.9% of patients had metastatic colorectal cancer. Patient payer status indicated that 36.1% were covered by Medicare; 25.6%, Medicare managed care; 4.3%, Medicaid; 30.3%, by commercial insurance and that 3.6% were uninsured or self-payers. As seen in the eAppendix Figure (eAppendix available at ajmc.com), the proportion of patients traveling more than 50 miles was low in all insurance groups. Finally, 52.7% of patients received surgery at a teaching hospital and 35.5% were operated upon by a general surgeon.
As shown in Table 2, compared with White patients, African American (odds ratio [OR], 0.29; 95% CI, 0.18-0.45) and Hispanic (OR, 0.13; 95% CI, 0.08-0.20) patients had reduced odds of traveling more than 50 miles for surgery. Patients located in urban areas also had reduced odds of traveling farther compared with those in rural areas (OR, 0.02; 95% CI, 0.01-0.03). Compared with patients with commercial insurance, those with Medicare (OR, 1.48; 95% CI, 1.03-2.15) had higher odds of traveling more than 50 miles for surgery. Similarly, patients living in South (OR, 2.37; 95% CI, 1.55-3.63) and Central (OR, 5.86; 95% CI, 3.87-8.89) Florida had higher odds of traveling for surgery compared with patients living in North Florida. Patients with metastasized colorectal cancer had increased odds of traveling farther (OR, 1.43; 95% CI, 1.12-1.82). Finally, hospital characteristics increasing the odds of travel included volume (OR, 1.12; 95% CI, 1.10-1.14), availability of colorectal surgeon (OR, 1.83; 95% CI, 1.45-2.30), and teaching status (OR, 9.96; 95% CI, 6.98-14.31).
The results of the analysis performed to assess bypassed vs chosen hospitals are revealed in Table 3. Hospitals chosen had increased odds of being medium sized vs large (OR, 3.08; 95% CI, 1.82-5.22) compared with the hospitals that patients bypassed. There was also a reduced odds of choosing a facility with lower colorectal surgical volume: 0 to 48 vs 150 to 277 surgeries (OR, 0.02; 95% CI, 0.01-0.03), 49 to 105 vs 150 to 277 surgeries (OR, 0.07; 95% CI, 0.04-0.11), and 106 to 149 vs 150 to 277 surgeries (OR, 0.37; 95% CI, 0.25-0.53). Next, patients had reduced odds of choosing a nonteaching hospital compared with a teaching hospital (OR, 0.08; 95% CI, 0.06-0.12). Finally, the chosen hospitals had greater odds of having a greater number of colorectal surgeons vs general surgeons (OR, 3.02; 95% CI, 2.21-4.12).
Table 4 shows the results of the multivariate model predicting the logarithmic transformation of LOS. Patients who traveled more than 50 miles to undergo surgery had a longer LOS compared with those who traveled less than 25 miles (β = 0.053; P = .03). Women had shorter LOS (β = –0.053; P < .001). LOS increased progressively with increasing age (β = 0.003; P < .001). African American patients had longer LOS compared with White patients (β = 0.044; P = .02). Patients with metastasis (β = 0.136; P < .001) and those with more comorbidities (1-2 vs 0: β = 0.109; P < .001; 3-5 vs 0: β = 0.347; P < .001; > 5 vs 0: β = 0.648; P < .001) had longer LOS. Compared with those with commercial insurance, patients with Medicaid and with self-pay/no insurance had longer LOS (β = 0.150; P < .001; and β = 0.127; P < .001, respectively). Patients who underwent surgical procedures at a teaching hospital had a longer LOS (β = 0.083; P < .001). In contrast, those at large and medium-sized hospitals had a shorter LOS (β = –0.193; P < .01; and β = –0.149; P < .001, respectively). Similarly, patients operated on by a colorectal surgeon had a shorter LOS (β = –0.077; P < .001).
As reported in Table 5, the multivariate model for surgical complications showed that traveling distance was not statistically different. Compared with open surgery, patients having minimally invasive surgery (MIS) had significantly lower odds of complication (OR, 0.76; 95% CI, 0.68-0.85). Patients receiving surgeries in teaching hospitals had higher odds of complication compared with those in nonteaching hospitals (OR, 1.43; 95% CI, 1.26-1.62). Patients with Medicaid (OR, 1.62; 95% CI, 1.24-2.11) or self-pay/no insurance (OR, 1.77; 95% CI, 1.34-2.35) had increased odds of complication. Similarly, female patients had reduced odds of complication compared with male patients (OR, 0.70; 95% CI, 0.63-0.78). Finally, the odds of complication significantly increased with a higher number of comorbidities.
The results of the sensitivity analysis to identify effect modifiers of the association between travel distance and LOS and between travel distance and complications are shown in eAppendix Tables 1 and 2. The effect modifiers of the association with LOS found in the analysis were Elixhauser score, payer group, and region of residence within Florida. Similarly, urbanization, obesity, and region of residence were identified to be effect modifiers of the association with complications. Stratified analysis revealed that within 1 or more strata of the identified effect modifiers, the OR/estimates of each outcome were significantly different from each other.
Our study focused on understanding what types of hospitals patients choose, as well as the influence of travel distance to surgical facility on the LOS and the rate of complications in patients with colorectal cancer. We found that patients with colorectal cancer who traveled more than 50 miles to a surgical facility were seen in medium-sized, higher-volume teaching hospitals that had colorectal surgeons. In addition, we found that those traveling further stayed longer at the surgical facility compared with those who traveled less than 25 miles. However, there were no differences in the rates of complication between the different travel groups.
Patients residing in the central and southern parts of Florida were observed to travel farther for surgery. Similarly, patients residing in rural areas traveled farther than those in urban areas. The discrepancies in travel distances to treatment may be attributable to centralization of care in the state of Florida.14 Further, this study found that patients who traveled more than 50 miles traveled to hospitals that had teaching status, had greater availability of a colorectal surgeon, and were medium-sized or high-volume centers. This finding is important because it helps identify the types of facilities that patients traveling farther are utilizing. It also provides opportunities to reduce travel time for patients by bolstering services in the hospitals that are closer to patients’ residences but are frequently bypassed.
The association of increased LOS at a treatment facility following an elective surgery with travel distance demonstrated in this study is consistent with that found in a previous study by Jackson et al.15 One possible explanation for this association is the type of surgical approach (open vs MIS) selected as the most appropriate for the elective surgery. It has been documented previously that the mean LOS following minimally invasive colectomy is significantly shorter compared with open colectomy (6.2 vs 8.7 days, respectively; P < .0001).16 In the current study, comparatively more patients selected an open surgical approach in the group who traveled farther, possibly explaining the longer LOS found in this group. Longer LOS may also be explained, to some extent, by additional barriers to timely discharge for patients who have to travel significant distances after surgery to reach their residence. Physicians may delay the discharge to ensure that these patients meet the predischarge requirements to avoid any adverse event during their long commute.17
Although significant discrepancies in the LOS were observed, our study did not find any significant difference in the rate of complications among the traveling groups. The underlying reason for no difference in the rate of complications, despite the worse metastatic status of the group that traveled farthest, is not well understood. One possible explanation could be the presence of inherent selection bias, as patients who traveled more than 50 miles for cancer care were younger and also had significantly lower comorbidity burden preoperatively. In a previous study, Etzioni et al found that patients who traveled lesser distance were at a lower risk of complication.18 Additionally, the risk of postsurgical complication is multifactorial. Patients’ baseline risk of having a complication, coupled with their ability to afford care to prevent the development of postoperative complications, should be taken into account when comparing the rate of complications between the travel groups.18 Further, the finding of higher rates of complications at teaching facilities should be interpreted with caution, as there were more patients at later stages of colorectal cancer traveling to these hospitals. The number of patients with colorectal cancer with metastasis was significantly higher in those who traveled more than 50 miles, which may indicate that these patients experienced a delayed diagnosis of colorectal cancer due to their distance from regional care centers and lack of easy access to colorectal cancer screening facilities.19
From a managed care perspective, patients with more advanced disease are usually best managed at tertiary centers.20 These facilities can provide care by more specialized professionals such as colorectal surgeons, and this is reflected in our analysis of characteristics of patients traveling further to receive care. However, in many cases, these patients are first seen in lower-volume hospitals and only subsequently referred to tertiary centers, which may lead to a delay in the definitive treatment that can potentially affect survival.21,22 Previous inquiry has identified colorectal surgical volume as an important consideration for valued outcomes, including procedural time, lower hospitalization cost, and shorter LOS.23 As such, the findings of the current study provide additional support toward provision of care for patients requiring colorectal surgery at higher-volume centers, as well as by colorectal surgeons. However, unlike previous studies, we found an increased LOS, which could increase financial costs, although this association may be due to procedural selections associated with caring for more advanced cancer stages. Although it is likely that better management of patients seeking colorectal cancer care could improve LOS, additional inquiry into assessing more granular data regarding patient LOS is warranted. Future efforts to improve the quality of patient outcomes in colorectal cancer should also focus on implementation of preventive care services that are better geographically distributed and accessible to diverse populations living in areas with a shortage of health care services.
This study has several limitations. Because the study was conducted using a discharge database, clinical information such as stage of cancer, blood loss, and operating time was not available. These factors could potentially be confounders of the association of travel distance with LOS or complications, but they could not be adjusted for in the multivariate regression models. The discharge database provided information based on a single admission and could not take into account complications that the patients might have experienced after discharge. We measured the distance from patient’s primary residence to surgical facility using zip code centroids under the assumption that all the patients used personal modes of transportation. In addition, although zip codes provide a fairly accurate point of measurement, the size of the zip code and the patient’s location within it are not accurately established nor adjusted for, which can either over- or underestimate distances to facilities. Our study also failed to adjust the travel distance for patients who used public transport, which may result in longer transportation times or other barriers to seeking care at facilities further away from their home.
We found that patients with colorectal cancer who traveled more than 50 miles to a surgical facility were more likely to be seen in medium-sized, higher-volume teaching hospitals that had colorectal surgeons. Additionally, patients with more advanced disease tended to travel more to receive treatment at a specialized center. We also found that those traveling more than 50 miles stayed longer at the surgical facility compared with those who traveled less than 25 miles to surgery. The current study, however, did not find any significant difference in the rates of complication among the different travel groups. Ultimately, the findings in this study may imply specialty regionalization in the state of Florida and a potential lack of preventive care services in areas distant from regional centers.
Author Affiliations: Department of Health Sciences Research (ACS, SB, JJC, NO) and Division of Colon and Rectal Surgery (RL, DTC), Mayo Clinic, Jacksonville, FL; Department of Medicine, Jacobi Medical Center (OO), New York, NY.
Source of Funding: None.
Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.
Authorship Information: Concept and design (ACS, SB, OO, JJC, DTC); acquisition of data (ACS, SB, JJC, NO, DTC); analysis and interpretation of data (ACS, SB, OO, JJC, RL, DTC); drafting of the manuscript (ACS, SB, OO, JJC, RL, NO, DTC); critical revision of the manuscript for important intellectual content (ACS, SB, OO, JJC, RL, NO, DTC); statistical analysis (ACS, SB, NO); administrative, technical, or logistic support (ACS, DTC); and supervision (ACS, DTC).
Address Correspondence to: Aaron C. Spaulding, PhD, Department of Health Sciences Research, Mayo Clinic, 4500 San Pablo Rd, Jacksonville, FL 32224. Email: email@example.com.
1. Zullig LL, Jackson GL, Provenzale D, Griffin JM, Phelan S, van Ryn M. Transportation: a vehicle or roadblock to cancer care for VA patients with colorectal cancer? Clin Colorectal Cancer. 2012;11(1):60-65. doi:10.1016/j.clcc.2011.05.001
2. Chou S, Deily ME, Li S. Travel distance and health outcomes for scheduled surgery. Med Care. 2014;52(3):250-257. doi:10.1097/MLR.0000000000000082
3. Murage P, Murchie P, Bachmann M, Crawford M, Jones A. Impact of travel time and rurality on presentation and outcomes of symptomatic colorectal cancer: a cross-sectional cohort study in primary care. Br J Gen Pract. 2017;67(660):e460-e466. doi:10.3399/bjgp17X691349
4. 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):2285-2293. doi:10.1007/s11605-019-04270-5
5. Wasif N, Chang YH, Pockaj BA, Gray RJ, Mathur A, Etzioni D. Association of distance traveled for surgery with short- and long-term cancer outcomes. Ann Surg Oncol. 2016;23(11):3444-3452. doi:10.1245/s10434-016-5242-z
6. Kelly M, Sharp L, Dwane F, Kelleher T, Comber H. Factors predicting hospital length-of-stay and readmission after colorectal resection: a population-based study of elective and emergency admissions. BMC Health Serv Res. 2012;12:77. doi:10.1186/1472-6963-12-77
7. Lassen K, Soop M, Nygren J, et al. Consensus review of optimal perioperative care in colorectal surgery: Enhanced Recovery After Surgery (ERAS) Group recommendations. Arch Surg. 2009;144(10):961-969. doi:10.1001/archsurg.2009.170
8. Harrison OJ, Smart NJ, White P, et al. Operative time and outcome of enhanced recovery after surgery after laparoscopic colorectal surgery. JSLS. 2014;18(2):265-272. doi:10.4293/108680813X13753907291918
9. Florida Center for Health Information and Policy Analysis. Patient databases general information. Updated March 2016. Accessed November 18, 2019. https://fhfstore.blob.core.windows.net/documents/researchers/OrderData/documents/KeyDifferences%20(Revised%202016).pdf
10. Clinical Classifications Software (CCS) for ICD-9-CM. Agency for Healthcare Research and Quality. Accessed December 5, 2019. https://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp
11. Patil CG, Alexander AL, Hayden Gephart MG, Lad SP, Arrigo RT, Boakye M. A population-based study of inpatient outcomes after operative management of nontraumatic intracerebral hemorrhage in the United States. World Neurosurg. 2012;78(6):640-645. doi:10.1016/j.wneu.2011.10.042
12. Onega T, Duell EJ, Shi X, Wang D, Demidenko E, Goodman D. Geographic access to cancer care in the U.S. Cancer. 2008;112(4):909-918. doi:10.1002/cncr.23229
13. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8-27. doi:10.1097/00005650-199801000-00004
14. Stitzenberg KB, Sigurdson ER, Egleston BL, Starkey RB, Meropol NJ. Centralization of cancer surgery: implications for patient access to optimal care. J Clin Oncol. 2009;27(28):4671-4678. doi:10.1200/JCO.2008.20.1715
15. Jackson KL, Glasgow RE, Hill BR, et al. Does travel distance influence length of stay in elective colorectal surgery? Dis Col Rectum. 2013;56(3):367-373. doi:10.1097/DCR.0b013e31827e939e
16. Bilimoria KY, Bentrem DJ, Merkow RP, et al. Laparoscopic-assisted vs. open colectomy for cancer: comparison of short-term outcomes from 121 hospitals. J Gastrointest Surg. 2008;12(11):2001-2009. doi:10.1007/s11605-008-0568-x
17. Agarwal S. Impact of travel time on same-day discharge after elective percutaneous coronary intervention. Tex Heart Inst J. 2017;44(6):431. doi:10.14503/THIJ-17-6467
18. Etzioni DA, Fowl RJ, Wasif N, Donohue JH, Cima RR. Distance bias and surgical outcomes. Med Care. 2013;51(3):238-244. doi:10.1097/MLR.0b013e318270bbfa
19. Ojinnaka CO, Choi Y, Kum HC, Bolin JN. Predictors of colorectal cancer screening: does rurality play a role? J Rural Health. 2015;31(3):254-268. doi:10.1111/jrh.12104
20. Tawfik Amin A, Salem AAS, Ibrahim A. Surgery for locally advanced GIT cancers has potentially good postoperative outcomes in a tertiary hospital. J Gastrointest Cancer. 2020;51(1):23-29. doi:10.1007/s12029-018-0181-5
21. Grass F, Behm KT, Duchalais E, et al. Impact of delay to surgery on survival in stage I-III colon cancer. Eur J Surg Oncol. 2020;46(3):455-461. doi:10.1016/j.ejso.2019.11.513
22. Flores-Balcázar CH, Flores-Luna ML, Villarreal-Garza CM, Bargalló-Rocha JE. Provider delay in treatment initiation and its influence on survival outcomes in women with operable breast cancer. Rep Pract Oncol Radiother. 2020;25(2):271-275. doi:10.1016/j.rpor.2020.02.002
23. Yi D, Monson JRT, Stankiewicz CC, Atallah S, Finkler NJ. Impact of colorectal surgeon case volume on outcomes and applications to quality improvement. Int J Colorectal Dis. 2018;33(5):635-644. doi:10.1007/s00384-018-3018-6