Association of Rurality, Race and Ethnicity, and Socioeconomic Status With the Surgical Management of Colon Cancer and Postoperative Outcomes Among Medicare Beneficiaries

Key Points Question Is rurality associated with receipt of surgery for nonmetastatic colon cancer? Findings In this cohort study of 57 710 Medicare beneficiaries with colon cancer, rural patients were more likely to undergo surgery but also more likely to receive emergent surgery and less likely to have minimally invasive surgery. Black race was independently associated with lower likelihood of surgery, and the association between rurality and postoperative mortality differed by race; rurality was protective against mortality for White beneficiaries, but rural-residing Black beneficiaries had increased postoperative mortality. Meaning Treatment and surgical outcomes in nonmetastatic colon cancer varied by rurality and by race and ethnicity, suggesting intersectional factors underlying disparities.


Introduction
2][3][4] Disparities in access to high-quality health care may be compounded for patients with cancer who require complex care from multiple clinicians. 4,50][11][12][13][14][15] Nonmetastatic colon cancer can be treated by general surgeons in most rural areas. 5However, prior descriptive studies focusing on geography 4,5,[9][10][11][12][13][14][15] have shown that patients living in rural areas with colon cancer are less likely to undergo cancer-directed surgical treatment and more likely to die of cancer than patients living in urban areas.Little is known about the interacting effects of rurality, race and ethnicity, and socioeconomic status on the treatment and outcomes of colon cancer. 16,17r objective was to evaluate the intersectionality of rurality, race and ethnicity, and socioeconomic status and the association of these characteristics with surgical outcomes in the treatment of nonmetastatic colon cancer.9][20] We hypothesized that patients living outside metropolitan areas would have less access to high-quality care, leading to inferior survival.We evaluated rurality in a national cohort of consistently insured patients for which we could separately control for race and ethnicity, socioeconomic status, and rurality.

Study Design and Cohort
We performed a cohort study of fee-for-service Medicare beneficiaries 65 years or older with incident, nonmetastatic colon cancer.This study was approved by the Committee for the Protection of Human Subjects at Dartmouth College, Lebanon, New Hampshire, which waived the need for informed consent because the research presented minimal risk of harm to participants and protections for privacy were in place, as required by the data use agreement.This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
We identified patients with incident colon cancer using a previously validated algorithm with minor modifications. 21,22We included beneficiaries meeting 1 or more of the following criteria: Among beneficiaries meeting at least 1 criterion for incident cancer, we excluded those with distant metastasis within 90 days of diagnosis (Figure 1).Evidence of distant metastasis was defined as any claim listing a diagnosis of a secondary malignant neoplasm (excepting the intra-abdominal lymph nodes).Because rectal cancer is sometimes miscoded as colon cancer in claims, we further excluded beneficiaries who had as many or more diagnoses of rectal cancer as for colon cancer within 60 days after index diagnosis (eMethods in the Supplement).The data that support the findings of this study are available with the permission from the Centers for Medicare & Medicaid Services (CMS).

Exposures and Outcomes
The primary exposure was rurality of patient residence according to zip code plus 4-digit extensions.Levels of rurality were defined using the US Department of Agriculture 2010 Rural-Urban Commuting Area (RUCA) codes and categorized as metropolitan (RUCA 1.0-3.0),micropolitan (RUCA 4.0-6.0),and small town or rural (RUCA 7.0-10.3).
Primary outcomes included (1) receipt of cancer-directed surgery (colectomy), (2) receipt of emergent surgery, (3) receipt of MIS, (4) 90-day postoperative surgical complications, and (5) 90-day postoperative mortality.Surgical resection was assessed in the full cohort, and the remaining surgical outcomes were assessed among beneficiaries undergoing surgery within 90 days of colon cancer diagnosis.
Emergent surgery, defined as surgery that was performed within 2 calendar days of an emergency department encounter or transfer from another facility, served as a measure of both access to care and overall care coordination. 18Minimally invasive surgery was defined as any laparoscopic or robotic procedure according to the procedure code from International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10), or Current Procedural Terminology.This outcome was selected as a marker of care quality because minimally invasive procedures have oncologic outcomes comparable to those for open procedures but have faster recovery times; accordingly, minimally invasive procedures are preferred when available. 20,23,24

Covariate Measures
All adjusted models accounted for age, sex, race and ethnicity, 27  raw score from 1 to 100, with higher scores indicating increased deprivation.The ADI was the only characteristic with missing data, missing for 2720 beneficiaries (4.7%), and we imputed the ADI for these beneficiaries (eMethods in the Supplement).Additional descriptive characteristics studied include hospital rurality, defined using the RUCA code of the hospital in which the surgery was performed, and the median distance traveled to surgical care, which was defined as the distance between zip code centroids for a patient's residence and the hospital where they underwent surgery.

Statistical Analysis
Data were analyzed from August 3, 2020, to April 30, 2021.We described differences in patient characteristics across rurality using proportions, means, and inferential statistics, including analysis of variance for continuous variables and χ 2 tests for categorical variables.
Adjusted analyses consisted of 2 stages.First, we used Kaplan-Meier failure curves, log-rank tests, and multivariable Cox proportional hazards regression to assess the association of rurality with time to surgery from the initial diagnosis.Patients were censored if they died or if they did not undergo surgery by the end of the study period.Second, we used multivariable logistic regression to examine the association between rurality and surgical outcomes among patients who underwent surgery within 90 days of diagnosis.All models were adjusted for the aforementioned covariates.
To study the intersectionality of rurality, race and ethnicity, and socioeconomic status with surgical management of colon cancer, we evaluated race and ethnicity, Medicare-Medicaid dual eligibility, and ADI as effect modifiers of the association between rurality and our primary outcomes using interaction terms.We performed an exploratory univariate analysis of the association between race and ethnicity with select key demographic characteristics and outcomes stratified by rurality.
We assessed statistical significance using the threshold of 2-tailed P < .05.All analyses were performed in Stata, version 16.1 (StataCorp LLC).

Results
We identified 57 710 Medicare beneficiaries diagnosed with incident, nonmetastatic colon cancer from April 1, 2016, to September 30, 2018.Among this group, 46.6% were men and 53.4% were women; the mean (SD) age was 76.6 (7.2) years.In terms of race and ethnicity, 3.7% were Hispanic, After adjustment, rurality was not associated with secondary outcomes of postoperative medical complications, receipt of adjuvant chemotherapy, or 90-day all-cause readmission (Table 3).
The exception was lower adjusted odds of postoperative medical complications for patients from small town and rural areas (OR, 0.89 [95% CI, 0.82-0.95])(Table 3).

Intersectionality of Rurality With Socioeconomic Status and Race and Ethnicity
Race and ethnicity constituted a modifier for the association of rurality with surgical complications (P = .001for interaction) and postoperative mortality (P = .001for interaction) but not for receipt of emergent surgery or MIS (Figure 2).The odds of a postoperative surgical complication were more than double for micropolitan-residing Hispanic patients (adjusted OR, 2. (1) any claim for inpatient or outpatient chemotherapy, radiotherapy, or cancer-directed surgery with a colon cancer diagnosis code; (2) 2 claims from different dates with a colon cancer diagnosis code within 12 months after a diagnostic biopsy; or (3) 2 claims with a colon cancer diagnosis code, with the first occurring within 14 days of a claim with a diagnosis code for a cancer-related symptom or complication and the second occurring within the 12 months after the cancer-related symptom.We JAMA Network Open | Public Health included patients with incident colon cancer diagnoses from April 1, 2016, to September 30, 2018, allowing for minimum follow-up of 90 days until December 31, 2018.

JAMA Network Open | Public Health Rurality
, Race and Ethnicity, and Socioeconomic Status in Surgical Management of Colon Cancer

Table 2 .
Estimated Primary Outcomes From Fully Adjusted Regression Models a Models were adjusted for age, sex, race and ethnicity, Medicare-Medicaid dual eligibility, ADI, cancer side, diabetes, CHF, COPD, previous MI, previous stroke or TIA, end-stage kidney disease, and CMS-HCC score. a

Table 3 .
Association of Rurality With Surgical Treatment and Outcomes Among Medicare Beneficiaries Undergoing Surgery Within 90 Days of Diagnosis for Incident Nonmetastatic Colon Cancer Rurality, Race and Ethnicity, and Socioeconomic Status in Surgical Management of Colon CancerPatients with Medicare-Medicaid dual eligibility in micropolitan areas and small town and rural areas were less likely to undergo minimally invasive procedures than non-dual eligible beneficiaries in the same level of rurality (P < .001forinteraction).Dual eligibility did not modify the association of rurality with emergent surgery or postoperative surgical complications (Figure2).2. Forest Plot of Adjusted Odds Ratios (ORs) for Primary Outcomes by Subgroups of Race and Ethnicity and Medicare-Medicaid Dual Eligibility Abbreviations: NA, not applicable; OR, odds ratio.a Models were adjusted for age, sex, race and ethnicity, Medicare-Medicaid dual eligibility, area deprivation index, cancer side, diabetes, congestive heart failure, chronic obstructive pulmonary disease, previous myocardial infarction, stroke or transient ischemic attack, end-stage kidney disease, and Centers for Medicare & Medicaid Services-Hierarchical Condition Category score.b Includes postprocedural hemorrhage or hematoma, wound infection, anastomotic leak, and abdominal abscess.c Includes kidney failure, pulmonary compromise, acute myocardial infarction, pneumonia, venous thrombosis and pulmonary embolism, and dementia or delirium.JAMA Network Open | Public Health JAMA Network Open.2022;5(8):e2229247.doi:10.1001/jamanetworkopen.2022.29247(Reprinted) August 30, 2022 8/14 Downloaded From: https://jamanetwork.com/ on 09/28/2023