Variation in Colorectal Cancer Stage and Mortality across Large Community-Based Populations: PORTAL Colorectal Cancer Cohort



 

Jennifer L Schneider, MPH1; Heather Spencer Feigelson, PhD, MPH2; Virginia P Quinn, PhD3;
Carmit McMullen, PhD4; Pamela A Pawloski, PharmD, BCOP, FCCP5; John D Powers, MS2; Andrew T Sterrett, PhD2;
David Arterburn, MD, MPH6; Douglas A Corley, MD, PhD1

Perm J 2020;24:19-182

https://doi.org/10.7812/TPP/19-182
E-pub: 07/21/2020

ABSTRACT

Introduction: Colorectal cancer (CRC) incidence and mortality can be reduced by effective screening and/or treatment. However, the influence of health care systems on disparities among insured patients is largely unexplored.

Methods: To evaluate insured patients with CRC diagnosed between 2010 and 2014 across 6 diverse US health care systems in the Patient-Centered Outcomes Research Institute (PCORI) Patient Outcomes Research To Advance Learning (PORTAL) CRC cohort, we contrasted CRC stage; CRC mortality; all-cause mortality; and influences of demographics, stage, comorbidities, and treatment between health systems.

Results: Among 16,211 patients with CRC, there were significant differences between health care systems in CRC stage at diagnosis, CRC-specific mortality, and all-cause mortality. The unadjusted risk of CRC mortality varied from 27% lower to 21% higher than the reference system (hazard ratio [HR] = 0.73, 95% confidence interval = 0.66-0.80 to HR = 1.21, 95% confidence interval = 1.05-1.40; p < 0.01 across systems). Significant differences persisted after adjustment for demographics and comorbidities (p < 0.01); however, adjustment for stage eliminated significant differences (p = 0.24). All-cause mortality among patients with CRC differed approximately 30% between health care systems (HR = 0.89-1.17; p < 0.01). Adjustment for age eliminated significant differences (p = 0.48).

Discussion: Differences in CRC survival between health care systems were largely explained by stage at diagnosis, not demographics, comorbidity, or treatment. Given that stage is strongly related to early detection, these results suggest that variation in CRC screening systems represents a modifiable systems-level factor for reducing disparities in CRC survival.

INTRODUCTION

Colorectal cancer (CRC) is the second leading cause of cancer death overall, but its incidence and mortality can be markedly reduced by effective screening, treatment, or both.1-3 The overall incidence of CRC in the US has decreased in recent years, mostly due to increases in CRC screening rates for recommended groups.4-6 However, disparities in both incidence and survival exist among certain races, age groups, and insurance types (access to care) and by cancer type (histology, morphology, genetic markers).7-10 For example, CRC incidence has actually increased among some subgroups, including younger persons (< 50 years old), and decreased more slowly among African Americans, Asians, and Hispanic whites than it has among non-Hispanic whites.11-13

The lack of studies evaluating systems-level differences in CRC outcomes has impeded addressing potentially modifiable disparities between health care systems in CRC outcomes. A few studies have suggested that access to care may diminish racial disparities in CRC outcomes, treatment, and recurrent cancers, although few studies have fully evaluated the entire process of care, including diagnosis, treatment, and survival.14-16 A challenge to studying variation in outcomes is a lack of studies comparing multiple health care systems with different underlying populations. Surveillance systems such as Surveillance, Epidemiology, and End Results (SEER) collect information on endpoints but lack risk factors such as comorbidities that may influence screening and treatment decisions or linkages to health care systems. In contrast, single medical centers may have rich data on a single population but lack variation in patient demographics, screening methods, or treatment patterns.17 To overcome these limitations, the Patient-Centered Outcomes Research Institute (PCORI) developed a scientific community and data resource of US patients, clinicians, and health care delivery systems. The Kaiser Permanente (KP) and Strategic Partners Patient Outcomes Research to Advance Learning (PORTAL) network is one of the PCORI-funded initiatives and includes a multisystem CRC cohort. In this study we leverage this cohort’s size—more than 16,000 patients—and its demographic diversity to estimate intersystem differences in postdiagnosis outcomes.18,19

The aim of this study was to identify major, potentially modifiable factors related to CRC mortality. in the current study we evaluated a large, diverse, multicenter cohort of patients with a diagnosis of CRC, and we contrasted CRC mortality and all-cause mortality across 6 distinct health care systems. Additionally, we evaluated whether patient-related factors (age, race/ethnicity, comorbidities) and health care system-related differences in stage-specific cancer treatments and in cancer stage (a surrogate for effective CRC screening) explained the differences found.

METHODS

Study Population

This observational cohort study was done in the PORTAL CRC cohort. Cohort development has been described elsewhere20; briefly, the cohort includes all adults (≥ 18 years of age) with CRC diagnosed between 2010 and 2014 from 6 health care systems. Those systems are Health Partners in Minneapolis, Minnesota; Kaiser Permanente Colorado (KPCO); Kaiser Permanente Northern California (KPNC); Group Health Cooperative in Washington State; Kaiser Permanente Northwest (KPNW); and Kaiser Permanente Southern California (KPSC). All patients are insured, although different insurance coverage exists in and across systems. Each system is administratively distinct in its practices for cancer screening and treatment.19 Health care systems are randomly designated A-F, with health care system A selected as the referent throughout. Colorectal cancer was defined using International Classification of Diseases O-3 codes C180, 182 to 189, C199, and C209. For this analysis, we excluded cohort members with SEER stage 0 disease, given inconsistent recording between cancer registries of carcinoma in situ.

Variables and Data Sources

Harmonized common data elements were available for all PORTAL health care systems. These data elements included demographics (age, race/ethnicity), comorbid conditions (Charlson comorbidity index), vital status (including cause of death, as appropriate), and social history. Cancer registry data characterized cancer diagnosis, location, treatment (surgery, chemotherapy, radiation therapy), tumor characteristics, additional patient demographic variables, and cause of death (supplemented by additional center-specific death registries).21

Statistical Analyses

Multivariable Cox proportional hazards regression models were used to calculate hazard ratios (HRs) as estimates of relative risk and 95% confidence intervals (CIs) for CRC mortality and all-cause mortality across health care systems. Health care system A served as the referent group. A sequential analytic approach, adding in factors individually or in related groups to evaluate their influence on model results, estimated the influence of demographics (age, sex, race/ethnicity, language preference, socioeconomic status) and comorbidities (abnormal body mass index [BMI], Charlson comorbidity index score)22-24; cancer stage (as a surrogate for system-specific factors such as screening that would influence early detection); tumor characteristics (morphology, grade); cancer treatments; and, as measures of opportunity to screen, health care system membership duration, year of diagnosis, and insurance coverage type.

The Cox proportional hazards assumption was evaluated by inspecting plots of the cumulative sums of Martingale residuals over follow-up times. All analyses were performed using statistical analysis software (SAS version 9.4, SAS Institute Inc, Cary, NC). The study was approved by the institutional review boards at all participating health care systems, and oversight was ceded to KPCO; informed consent was not required. In all models, patients were censored at the first of the following events: end of health care system membership, death (non-CRC death in CRC death analyses), or end of the study period (December 31, 2014).

Given strong known associations between screening, stage, and CRC mortality 2-6, logistic regression was used to evaluate risk factors, independent of health care system, for late-stage (stages 3 and 4) vs early-stage disease (stages 1 and 2). These supplemental analyses allowed evaluation of whether between-system differences were independent of variation in, for example, age, race/ethnicity, or comorbidities between systems. 

RESULTS

Cohort

The CRC cohort characteristics have been previously described.20 A total of 16,211 persons with diagnosed CRC were identified. Exclusion of 1539 persons with stage 0 cancer provided 14,672 persons for the main analyses. The racial/ethnic distribution was 64% non-Hispanic white, 15% Hispanic, 11% Asian/Pacific Islander, and 10% African American. Half were female (50%); most were greater than 65 years old (mean age at diagnosis = 68 years); and most had commercial insurance alone or in combination with another insurance type (66%; Table 1). The number of CRC diagnoses per year was similar over the study period; 89% of cancers were nonmucinous adenocarcinomas, and 64% had moderately differentiated histology. The most common initial treatments were surgery (85%), chemotherapy (39%), radiation therapy (11%), and palliative care (< 1%); some patients received more than 1 initial treatment type. The mean follow-up after CRC diagnosis was 1.94 (SD = 1.42) years (range by site = 1.73-1.98 years). During follow-up, 1179 patients were censored at disenrollment from the health care system, and 3624 patients died of any cause, among whom 2415 had CRC-related deaths.

Colorectal Cancer-Specific Mortality

The HRs for CRC-specific mortality varied substantially (> 45%) and significantly across health care systems in unadjusted models (HR range referent to health care system A = 0.73-1.21, p < 0.01). Significant variation of approximately 40% remained after adjustment for potential differences in demographics (age group, race/ethnicity, sex) and health status (Charlson comorbidity index score and BMI) between health systems (p=.0002). In contrast, adjustment for stage, which is a surrogate of screening/early detection, eliminated significant differences between health care systems (p = 0.24, Figure 1).

Additional evaluations of the influences of tumor characteristics, insurance type, receipt of treatment, and year of diagnosis did not further influence differences in CRC mortality.

All-Cause Mortality

The HRs for all-cause mortality among patients with CRC differed approximately 30% between health care systems (HR range = 0.89-1.17, p < 0.01); however, adjustment for age alone eliminated significant differences (p = 0.48). Adjustment for both demographics and comorbidity provided even more comparable estimates (Figure 2). The additional inclusion of BMI, co-morbid conditions, and stage at diagnosis did not substantially further influence variability across health care systems in all-cause mortality (p = 0.50).

Early versus Late Stage at Diagnosis

Given that cancer stage, as a surrogate of screening and early detection, is a primary driver of CRC mortality, we evaluated predictors of CRC stage at diagnosis among the 14,224 patients with stages 1 to 4 disease (late stage [3/4] vs early stage [1/2] at diagnosis, Figure 3). Significant variation in late-stage vs early-stage disease across systems was observed after adjustment for demographic factors, health status, health care system membership duration, and insurance type (HR range = 0.77-1.19, p = 0.03).

Factors associated with later-stage disease, independent of health care system, included African American race (odds ratio [OR] = 1.12, 95% CI = 1.00-1.26), age outside usual screening intervals (ie, < 50 or > 80 years), and greater numbers of comorbidities (eg, Charlson comorbidity index score 3: OR = 1.36, 95% CI = 1.28-1.43; Table 2).

DISCUSSION

In this cohort study of 6 US health care systems, using the PORTAL CRC cohort, we found substantial and significant differences between health care systems in CRC-specific mortality, all-cause mortality, and stage of CRC at diagnosis. The between-system differences in all-cause mortality were largely explained by differing age structures between populations. In contrast, the between-system differences in CRC-specific mortality were explained almost solely by differences in cancer stage, a surrogate for effective screening, and not by differences in demographics, comorbidities, or cancer treatment. Combined with the observation that the CRC-associated mortality was higher among persons older or younger than recommended screening ages, these findings suggest that systems-level factors in early detection (ie, CRC screening programs) likely explain the demonstrated disparities in CRC mortality between health care systems.

Few studies have contrasted systems-level data on CRC outcomes; thus, the current study informs the relative roles that demographics, stage/early detection, and treatments may have in influencing CRC-specific and all-cause mortality among patients with a diagnosis of CRC. Prior studies demonstrated substantial variation in screening completion, even in integrated health systems, although these were not linked with outcomes data.25

Race is one of the most commonly studied factors related to variation in CRC outcomes. Several races, in general, have demonstrated poorer outcomes than whites in some settings, although a small increased risk of advanced-stage disease in the current study was found only among African Americans (OR = 1.12, 95% CI = 1.00-1.26) and not among other race/ethnicity groups compared with whites. The current findings are consistent with those of another recent study, which found disparities in CRC outcomes by race in nonintegrated health care settings but not in integrated health care settings.26 The current study now finds that even between integrated systems, race does not explain disparities in CRC survival; rather, it is related to stage, independent of race.

Although we were not able to directly assess variations in screening methods and proportions screened across systems, among persons with access to care, screening is the main known cause of early detection; thus stage of cancer at diagnosis serves as a likely marker for penetrance of the screening program in a health care system. Screening programs that are well accepted and broadly implemented will detect cancers at earlier stages than programs with less coverage and can directly and markedly reduce CRC mortality.27-30 In this study, health care system F had lower risks of both CRC-specific mortality and late-stage diagnosis (OR = 0.89, 95% CI = 0.83-0.96). These results are concordant with preexisting knowledge, from prior analyses, of differences in screening rates and follow-up of abnormal screening tests between some of the programs under evaluation.31-34 In addition, patients outside screening ages (50-75 years for average risk), as expected, also were more likely to have a late-stage diagnosis. Screening differences between health care systems can include variation in formal outreach and differences in use of the most effective tests, such as colonoscopy or fecal immunochemical testing, vs use of less effective tests, such as sigmoidoscopy or fecal occult blood testing.35

Further exploration into the differences in screening practices of individuals across health systems would better inform modifiable differences between health care systems. Well-designed screening programs with high overall proportions of persons who are up to date with screening can have variable uptake by different patient populations. Even with relatively standardized screening offerings, the uptake, follow-up of patients with abnormal test results, and the screening test choices can vary by race, age, and comorbidity status.23,29 Access to care and insurance coverage are known to vary by demographic group and can affect severity of disease and eventually mortality; however, our study design adjusted for this by only evaluating persons with health care insurance coverage.36

There are several strengths to this study. The multicenter cohort is one of the largest, most diverse CRC cohorts with individual-level data yet described. The demographic and geographic diversity (15% Hispanic, 10% African American, and 11% Asian/Pacific Islander; almost 1500 persons who received a diagnosis under the age of 50 years) allow analyses of multiple factors that may influence CRC survival and all-cause mortality across 6 distinct, large health care systems with different care delivery models. Limitations include cohort homogeneity regarding access to care, although this can also be a strength, because this homogeneity effectively “controlled” for health care access, allowing less biased evaluations of other factors. As noted, although the study assessed stage, it did not directly assess screening completion. Thus, the analyses could not evaluate how variations in screening rates influence stage independent of race, sex, age, and other factors.

CONCLUSION

Among patients with a diagnosis of CRC cancer, there is substantial and significant variation in both CRC-related mortality and all-cause mortality across different health care systems. Although the differences in all-cause mortality were largely explained by differences in age between the health care systems, the differences in CRC-related mortality were largely explained by differences in cancer stage, not by differences in demographics or cancer treatment. Given that the main determinant of cancer stage is the use of effective CRC screening tests for early detection, and the demonstrated prior differences in screening and screening follow-up rates between some of the centers evaluated, these results suggest that more consistent application of cancer screening across health care systems may further reduce the current disparities in CRC mortality between health care systems.

Disclosure Statement

The author(s) have no conflicts of interest to disclose.

Acknowledgments

This study used infrastructure developed by the PORTAL (Patient Outcomes Research to Advance Learning) Network, a consortium of 3 integrated delivery systems (Kaiser Permanente, HealthPartners, and Denver Health) and their affiliated research centers. The PORTAL Network also performed data collection.

Research reported in this article was funded through Patient-Centered Outcomes Research Institute (PCORI) Award CDRN-1306-04681 Phase II. The statements in this article are solely the responsibility of the authors and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute (PCORI), its Board of Governors, or Methodology Committee.

Kathleen Louden, ELS, of Louden Health Communications performed a primary copyedit.

Author Affiliations

1 Division of Research, Kaiser Permanente Northern California, Oakland
2 Institute for Health Research, Kaiser Permanente Colorado, Denver
3 Department of Research and Evaluation, Kaiser Permanente Southern California, Pasadena
4 Center for Health Research, Kaiser Permanente Northwest, Portland, OR
5 HealthPartners Institute, Bloomington, MN
6 Kaiser Permanente Washington Health Research Institute, Seattle

Corresponding Author

Jennifer Schneider, MPH ()

Author Contributions

Heather Spencer Feigelson, PhD, MPH, and Douglas A. Corley, MD, PhD, assisted with study design and data analysis. John D. Powers, MS, participated in data analysis. Jennifer L. Schneider, MPH, drafted the manuscript. All authors reviewed, edited, and approved the final manuscript.

References

1.    Aran, V., et al., Colorectal Cancer: Epidemiology, Disease Mechanisms and Interventions to Reduce Onset and Mortality. Clin Colorectal Cancer, 2016. 15(3): p. 195-203 DOI: 10.1016/j.clcc.2016.02.008.
    2.    Zauber, A.G., The impact of screening on colorectal cancer mortality and incidence: has it really made a difference? Dig Dis Sci, 2015. 60(3): p. 681-91 DOI: 10.1007/s10620-015-3600-5.
    3.    Nishihara, R., et al., Long-term colorectal-cancer incidence and mortality after lower endoscopy. N Engl J Med, 2013. 369(12): p. 1095-105 DOI: 10.1056/NEJMoa1301969.
    4.    Austin, H., et al., Changes in colorectal cancer incidence rates in young and older adults in the United States: what does it tell us about screening. Cancer Causes Control, 2014. 25(2): p. 191-201 DOI: 10.1007/s10552-013-0321-y.
    5.    Levin, T.R., et al., Effects of Organized Colorectal Cancer Screening on Cancer Incidence and Mortality in a Large Community-Based Population. Gastroenterology, 2018. 155(5): p. 1383-1391 e5 DOI: 10.1053/j.gastro.2018.07.017.
    6.    Doubeni, C.A., et al., Modifiable Failures in the Colorectal Cancer Screening Process and Their Association with Risk of Death. Gastroenterology, 2018 DOI: 10.1053/j.gastro.2018.09.040.
    7.    DeSantis, C.E., et al., Cancer statistics for African Americans, 2016: Progress and opportunities in reducing racial disparities. CA Cancer J Clin, 2016. 66(4): p. 290-308 DOI: 10.3322/caac.21340.
    8.    Alexander, D.D., et al., African-American and Caucasian disparities in colorectal cancer mortality and survival by data source: an epidemiologic review. Cancer Biomark, 2007. 3(6): p. 301-13.
    9.    Pardini, B., et al., Polymorphisms in microRNA genes as predictors of clinical outcomes in colorectal cancer patients. Carcinogenesis, 2015. 36(1): p. 82-6 DOI: 10.1093/carcin/bgu224.
    10.    Phipps, A.I., et al., Common genetic variation and survival after colorectal cancer diagnosis: a genome-wide analysis. Carcinogenesis, 2016. 37(1): p. 87-95 DOI: 10.1093/carcin/bgv161.
    11.    O’Connell, J.B., et al., Rates of colon and rectal cancers are increasing in young adults. Am Surg, 2003. 69(10): p. 866-72.
    12.    Ashktorab, H., et al., Colorectal Cancer in Young African Americans: Is It Time to Revisit Guidelines and Prevention? Dig Dis Sci, 2016. 61(10): p. 3026-30 DOI: 10.1007/s10620-016-4207-1.
    13.    Siegel, R.L., et al., Colorectal Cancer Incidence Patterns in the United States, 1974-2013. J Natl Cancer Inst, 2017. 109(8) DOI: 10.1093/jnci/djw322.
    14.    Haider, A.H., et al., Racial disparities in surgical care and outcomes in the United States: a comprehensive review of patient, provider, and systemic factors. J Am Coll Surg, 2013. 216(3): p. 482-92.e12 DOI: 10.1016/j.jamcollsurg.2012.11.014.
    15.    Laiyemo, A.O., et al., Race and colorectal cancer disparities: health-care utilization vs different cancer susceptibilities. J Natl Cancer Inst, 2010. 102(8): p. 538-46 DOI: 10.1093/jnci/djq068.
    16.    Rutter, M.D., et al., World Endoscopy Organization Consensus Statements on Post-Colonoscopy and Post-Imaging Colorectal Cancer. Gastroenterology, 2018 DOI: 10.1053/j.gastro.2018.05.038.
    17.    Levin, T.R., et al., Effects of Organized Colorectal Cancer Screening on Cancer Incidence and Mortality in a Large, Community-based Population. Gastroenterology, 2018 DOI: 10.1053/j.gastro.2018.07.017.
    18.    Corley, D.A., et al., Building Data Infrastructure to Evaluate and Improve Quality: PCORnet. J Oncol Pract, 2015. 11(3): p. 204-6 DOI: 10.1200/jop.2014.003194.
    19.    McGlynn, E.A., et al., Developing a data infrastructure for a learning health system: the PORTAL network. J Am Med Inform Assoc, 2014. 21(4): p. 596-601 DOI: 10.1136/amiajnl-2014-002746.
    20.    Feigelson, H.S., et al., Optimizing patient-reported outcome and risk factor reporting from cancer survivors: a randomized trial of four different survey methods among colorectal cancer survivors. J Cancer Surviv, 2017. 11(3): p. 393-400 DOI: 10.1007/s11764-017-0596-1.
    21.    North American Association of Central Cancer Registries. 2018 [cited 2018 Nov 29]. Available from: https://www.naaccr.org/cina-data-products-overview/.
    22.    Klabunde, C.N., et al., A refined comorbidity measurement algorithm for claims-based studies of breast, prostate, colorectal, and lung cancer patients. Ann Epidemiol, 2007. 17(8): p. 584-90 DOI: 10.1016/j.annepidem.2007.03.011.
    23.    Klabunde, C.N., et al., Influence of Age and Comorbidity on Colorectal Cancer Screening in the Elderly. Am J Prev Med, 2016. 51(3): p. e67-75 DOI: 10.1016/j.amepre.2016.04.018.
    24.    Cho, H., et al., Comorbidity-adjusted life expectancy: a new tool to inform recommendations for optimal screening strategies. Ann Intern Med, 2013. 159(10): p. 667-76 DOI: 10.7326/0003-4819-159-10-201311190-00005.
    25.    Green, B.B., et al., A centralized mailed program with stepped increases of support increases time in compliance with colorectal cancer screening guidelines over 5 years: A randomized trial. Cancer, 2017. 123(22): p. 4472-4480 DOI: 10.1002/cncr.30908.
    26.    Rhoads, K.F., et al., How do integrated health care systems address racial and ethnic disparities in colon cancer? J Clin Oncol, 2015. 33(8): p. 854-60 DOI: 10.1200/jco.2014.56.8642.
    27.    Doubeni, C.A., et al., Effectiveness of screening colonoscopy in reducing the risk of death from right and left colon cancer: a large community-based study. Gut, 2018. 67(2): p. 291-298 DOI: 10.1136/gutjnl-2016-312712.
    28.    Cole, S.R., et al., Shift to earlier stage at diagnosis as a consequence of the National Bowel Cancer Screening Program. Med J Aust, 2013. 198(6): p. 327-30.

    29.    Fedewa, S.A., et al., Colorectal Cancer Screening Initiation After Age 50 Years in an Organized Program. Am J Prev Med, 2017. 53(3): p. 335-344 DOI: 10.1016/j.amepre.2017.02.018.
    30.    Knudsen, A.B., et al., Estimation of Benefits, Burden, and Harms of Colorectal Cancer Screening Strategies: Modeling Study for the US Preventive Services Task Force. JAMA, 2016. 315(23): p. 2595-609 DOI: 10.1001/jama.2016.6828.
    31.    Tosteson, A.N., et al., Variation in Screening Abnormality Rates and Follow-Up of Breast, Cervical and Colorectal Cancer Screening within the PROSPR Consortium. J Gen Intern Med, 2016. 31(4): p. 372-9 DOI: 10.1007/s11606-015-3552-7.
    32.    McCarthy, A.M., et al., Follow-Up of Abnormal Breast and Colorectal Cancer Screening by Race/Ethnicity. Am J Prev Med, 2016. 51(4): p. 507-12 DOI: 10.1016/j.amepre.2016.03.017.
    33.    Burnett-Hartman, A.N., et al., Racial/Ethnic Disparities in Colorectal Cancer Screening Across Healthcare Systems. Am J Prev Med, 2016. 51(4): p. e107-15 DOI: 10.1016/j.amepre.2016.02.025.
    34.    Chubak, J., et al., Time to Colonoscopy after Positive Fecal Blood Test in Four U.S. Health Care Systems. Cancer Epidemiol Biomarkers Prev, 2016. 25(2): p. 344-50 DOI: 10.1158/1055-9965.EPI-15-0470.
    35.    Bibbins-Domingo, K., et al., Screening for Colorectal Cancer: US Preventive Services Task Force Recommendation Statement. Jama, 2016. 315(23): p. 2564-2575 DOI: 10.1001/jama.2016.5989.
    36.    Grant, S.R., et al., Variation in insurance status by patient demographics and tumor site among nonelderly adult patients with cancer. Cancer, 2015. 121(12): p. 2020-8 DOI: 10.1002/cncr.29120.

Keywords: Colorectal cancer, health care systems, mortality, patient outcomes research, variation

 

Table 1. Descriptive characteristics of patients with colorectal cancer (CRC), treatment modalities, and cancer outcomes for PORTAL CRC cohort, by health care system (N = 14,672)a
Characteristic System A, no. (%) System B, no. (%) System C no. (%) System D no. (%) System E no. (%) System F no. (%) Total
no. (%)
Number (%) 6073 (41.4) 1020 (7.0) 180 (1.2) 860 (5.9) 935 (6.4) 5604 (38.2) 14,672 (100.0)
Sex
Female 3070 (50.6) 540 (52.9) 98 (54.4) 462 (53.7) 462 (49.4) 2652 (47.3) 7284 (49.6)
Male 3003 (49.4) 480 (47.1) 82 (45.6) 398 (46.3) 473 (50.6) 2952 (52.7) 7388 (50.4)
Race/ethnicity
African American 531 (8.7) 34 (3.3) 9 (5.0) 36 (4.2) 26 (2.8) 752 (13.4) 1388 (9.5)
Asian/Pacific Islander 842 (13.9) 75 (7.4) 9 (5.0) 22 (2.6) 28 (3.0) 572 (10.2) 1548 (10.6)
Hispanic 766 (12.6) 24 (2.4) 1 (0.6) 88 (10.2) 28 (3.0) 1323 (23.6) 2230 (15.2)
White 3904 (64.3) 862 (84.5) 160 (88.9) 647 (75.2) 844 (90.3) 2923 (52.2) 9340 (63.7)
Other/unknown 30 (0.5) 25 (2.5) 1 (0.6) 67 (7.8) 9 (1.0) 34 (0.6) 166 (1.1)
Age, y, at CRC Dx
< 40 146 (2.4) 17 (1.7) 2 (1.1) 16 (1.9) 14 (1.5) 163 (2.9) 358 (2.4)
40-49 473 (7.8) 57 (5.6) 16 (8.9) 60 (7.0) 55 (5.9) 465 (8.3) 1126 (7.7)
50-59 1183 (19.5) 176 (17.3) 32 (17.8) 165 (19.2) 166 (17.8) 1171 (20.9) 2893 (19.7)
60-69 1476 (24.3) 260 (25.5) 40 (22.2) 215 (25.0) 286 (30.6) 1491 (26.6) 3768 (25.7)
70-79 1433 (23.6) 231 (22.6) 38 (21.1) 205 (23.8) 217 (23.2) 1312 (23.4) 3436 (23.4)
≥ 80 1362 (22.4) 279 (27.4) 52 (28.9) 199 (23.1) 197 (21.1) 1002 (17.9) 3091 (21.1)
Mean (SD) 67.83 (13.96) 69.77 (13.6) 69.33 (14.49) 68.32 (13.35) 68.28 (12.87) 66.52 (13.55) 67.54 (13.74)
BMI, kg/m2, mean (SD) 27.67 (6.29) 28.49 (6.5) 28.6 (7.05) 27.84 (6.23) 29.65 (7.1) 28.1 (6.17) 28.03 (6.34)
Charlson comorbidity index score
1 1122 (18.5) 65 (6.4) 7 (3.9) 125 (14.5) 61 (6.5) 1557 (27.8) 2937 (20.0)
2.1 578 (9.5) 46 (4.5) 5 (2.8) 50 (5.8) 32 (3.4) 703 (12.5) 1414 (9.6)
3.2 1195 (19.7) 260 (25.5) 54 (30.0) 213 (24.8) 309 (33.0) 1028 (18.3) 3059 (20.8)
≥ 4.3 3178 (52.3) 649 (63.6) 114 (63.3) 472 (54.9) 533 (57.0) 2316 (41.3) 7262 (49.5)
Enrollment duration before Dx, y, mean (SD) 3.77 (1.74) 3.7 (0.02) 3.56 (0.02) 3.49 (0.02) 3.57 (1.8) 3.62 (1.9) 3.67 (1.82)
Year of Dx
2010 1326 (21.8) 187 (18.3) 36 (20.0) 187 (21.7) 211 (22.6) 1128 (20.1) 3075 (21.0)
2011 1317 (21.7) 192 (18.8) 45 (25.0) 161 (18.7) 197 (21.1) 1081 (19.3) 2993 (20.4)
2012 1201 (19.8) 208 (20.4) 41 (22.8) 193 (22.4) 183 (19.6) 1134 (20.2) 2960 (20.2)
2013 1228 (20.2) 227 (22.3) 44 (24.4) 187 (21.7) 176 (18.8) 1145 (20.4) 3007 (20.5)
2014 1001 (16.5) 206 (20.2) 14 (7.8) 132 (15.3) 168 (18.0) 1116 (19.9) 2637 (18.0)
Stage at Dx
1 1646 (27.1) 239 (23.4) 60 (33.3) 250 (29.1) 263 (28.1) 1702 (30.4) 4160 (28.4)
2 1555 (25.6) 266 (26.1) 48 (26.7) 202 (23.5) 242 (25.9) 1413 (25.2) 3726 (25.4)
3 1596 (26.3) 272 (26.7) 46 (25.6) 222 (25.8) 213 (22.8) 1417 (25.3) 3766 (25.7)
4 1117 (18.4) 194 (19.0) 21 (11.7) 159 (18.5) 144 (15.4) 937 (16.7) 2572 (17.5)
Unknown 159 (2.6) 49 (4.8) 5 (2.8) 27 (3.1) 73 (7.8) 135 (2.4) 448 (3.1)
Morphology
Nonmucinous adenocarcinoma 5472 (90.1) 901 (88.3) 159 (88.3) 769 (89.4) 839 (89.7) 4934 (88.0) 13074 (89.1)
Mucinous adenocarcinoma 389 (6.4) 71 (7.0) 15 (8.3) 61 (7.1) 69 (7.4) 495 (8.8) 1100 (7.5)
Signet ring 64 (1.1) 10 (1.0) 2 (1.1) 7 (0.8) 9 (1.0) 91 (1.6) 183 (1.2)
Other/NOS 148 (2.4) 38 (3.7) 4 (2.2) 23 (2.7) 18 (1.9) 84 (1.5) 315 (2.1)
Histology
Well differentiated 393 (6.5) 54 (5.3) 7 (3.9) 65 (7.6) 126 (13.5) 651 (11.6) 1296 (8.8)
Moderately differentiated 4118 (67.8) 642 (62.9) 121 (67.2) 514 (59.8) 540 (57.8) 3511 (62.7) 9446 (64.4)
Poorly differentiated 661 (10.9) 130 (12.7) 37 (20.6) 145 (16.9) 133 (14.2) 967 (17.3) 2073 (14.1)
Undifferentiated 155 (2.6) 38 (3.7) 1 (0.6) 53 (6.2) 7 (0.7) 53 (0.9) 307 (2.1)
Not determined 746 (12.3) 156 (15.3) 14 (7.8) 83 (9.7) 129 (13.8) 422 (7.5) 1550 (10.6)
Treatment
Surgery 5092 (83.8) 868 (85.1) 154 (85.6) 726 (84.4) 796 (85.1) 4779 (85.3) 12415 (84.6)
Chemotherapy 2312 (38.1) 395 (38.7) 65 (36.1) 329 (38.3) 318 (34.0) 2367 (42.2) 5786 (39.4)
Radiation therapy 594 (9.8) 147 (14.4) 25 (13.9) 123 (14.3) 109 (11.7) 681 (12.2) 1679 (11.4)
Palliative care 0 (0) 0 (0) 23 (12.8) 33 (3.8) 72 (7.7) 0 (0) 128 (0.9)
Insurance type
Medicaid 77 (1.3) 0 (0) 15 (8.3) 31 (3.6) 14 (1.5) 75 (1.3) 212 (1.4)
Medicare and private pay 1845 (30.4) 391 (38.3) 0 (0) 187 (21.7) 0 (0) 1722 (30.7) 4145 (28.3)
Medicare and commercial 1741 (28.7) 259 (25.4) 0 (0) 225 (26.2) 35 (3.7) 25 (0.5) 2285 (15.6)
High-deductible 137 (2.3) 35 (3.4) 12 (6.7) 18 (2.1) 1 (0.1) 401 (7.2) 604 (4.1)
Commercial 2273 (37.4) 335 (32.8) 153 (85) 399 (46.4) 885 (94.7) 3381 (60.3) 7426 (50.6)
Died of any cause 1582 (25.1) 279 (27.4) 47 (26.1) 223 (25.9) 226 (24.2) 1267 (22.6) 3624 (24.7)
CRC death 1141 (19.0) 212 (20.9) 24 (14.5) 159 (18.8) 167 (17.9) 712 (13.7) 2415 (17.0)
Follow-up time, y, mean (SD) 1.98 (1.42) 1.73 (1.37) 1.96 (1.33) 1.94 (1.41) 1.93 (1.44) 1.93 (1.42) 1.94 (1.42)

a Some totals do not total to 100% because of rounding.
BMI = body mass index; Dx = diagnosis; NOS = not otherwise specified; PORTAL = Patient Outcomes Research To Advance Learning; SD = standard deviation.

 

Table 2. Individual predictors of stage at diagnosis (N = 14,224)
Predictor Hazard ratio (95% CI)
Body mass index 0.98 (0.97-0.98)
Race/ethnicity (referent = white non-Hispanic)
African American 1.12 (1.00-1.26)
Asian/Pacific Islander 0.97 (0.87-1.08)
Hispanic 0.99 (0.89-1.09)
Other/unknown 0.94 (0.72-1.21)
Age group, y (referent = 50-59)
< 40 2.02 (1.67-2.44)
40-49 1.59 (1.42-1.78)
50-59 1 (referent)
60-69 0.75 (0.70-0.80)
70-79 0.68 (0.62-0.74)
≥ 80 0.70 (0.64-0.77)
Sex (referent = male) 1 (0.97-1.03)
Years enrolled 0.97 (0.96-0.99)
Charlson comorbidity index score (referent = 0)
1 0.95 (0.87-1.04)
2 0.89 (0.83-0.95)
≥ 3 1.36 (1.28-1.43)
Insurance type (referent = commercial only)
Medicaid 1.04 (0.82-1.30)
Medicare and private pay 0.95 (0.86-1.04)
Medicare and commercial 0.99 (0.89-1.11)
High-deductible 1.04 (0.90-1.21)

CI = confidence interval.

 

19 182 figure 1

19 182 figure 2

19 182 figure 3

ETOC

Click here to join the eTOC list or text ETOC to 22828. You will receive an email notice with the Table of Contents of The Permanente Journal.

CIRCULATION

2 million page views of TPJ articles in PubMed from a broad international readership.

Indexing

Indexed in MEDLINE, PubMed Central, HINARI, EMBASE, EBSCO Academic Search Complete, rdrb, CrossRef, and SciVerse/Scopus.


                                             

 

 

ISSN 1552-5767 Copyright © 2020 thepermanentejournal.org.

All Rights Reserved.