Gender and Ethnic Disparities in Colon Cancer Presentation and Outcomes in a US Universal Health Care Setting

Transcription

1 Gender and Ethnic Disparities in Colon Cancer Presentation and Outcomes in a US Universal Health Care Setting RAMZI AMRI, MSc, 1 KARIEN STRONKS, PhD, 2 LILIANA G. BORDEIANOU, MD, 1 PATRICIA SYLLA, MD, 1 AND DAVID L. BERGER, MD 1 * 1 Division of General and Gastrointestinal Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 2 Department of Public Health, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands Objective: Access to care is a pillar of U.S. healthcare reform and could potentially challenge existing ethnic and gender disparities in care. We present a snapshot of these disparities in surgical colon cancer patients in the largest public hospital in Massachusetts, a state leading in providing universal healthcare, to indicate potential changes that might result from universal care access. Methods: All surgical colon cancer patients at Massachusetts General Hospital ( ) were included. Baseline characteristics, perioperative, and long term outcomes were compared. Results: Among 1,071 patients, the 110 (10.3%) minority patients presented with more comorbid (mean Charlson score 0.84 vs. 0.71; P ¼ 0.039), metastatic (21.8% vs. 14%; P ¼ 0.026), and node positive disease (50% vs. 38.8%; P ¼ 0.014). Women (n ¼ 521; 48.6%) had less screening diagnoses (overall: 17.8% vs. 22.6%; P ¼ 0.049, screening age: 26.4% vs. 32.7%; P ¼ 0.036) with subsequently higher rates of metastatic disease on pathology (11.3% vs. 7.1%, P ¼ 0.02). Multivariate adjustment for baseline staging makes outcome disparities no longer statistically significant. Conclusions: Significant gender and ethnic disparities subsist at baseline despite long standing low threshold healthcare access, although seemingly mitigated by enrollment into high level care, empowering equal chances for underprivileged groups. The outcomes are also a reminder that universal healthcare will not be a panacea for the deeply rooted and dynamic causes of presentation inequalities. J. Surg. Oncol. ß 2014 Wiley Periodicals, Inc. KEY WORDS: colonic neoplasms; healthcare disparities; minority health; health status disparities; neoplasm staging; disease free survival; mortality INTRODUCTION Healthcare disparities between genders and ethnic groups are widespread in the United States [1,2] and are thought to be interrelated [3]. These disparities relate to complex intertwined combinations of factors that include higher morbidity and disease mortality rates [4], lower health literacy [5], lower quality of provided care [6], and worse access to healthcare [7]. The impact of these factors is best illustrated by the life expectancy gap of over 4.5 years between Whites and African Americans [8]. Colon cancer is a disease of the elderly [9] with associations with several lifestyle factors, including smoking [10], diabetes [11], and obesity [12]. Colorectal cancer has a comparable incidence across ethnicities and sexes and is second only to lung cancer in terms of yearly deaths in the US, with colon cancer leading to over two thirds of these deaths [13]. Ethnic disparities, with more advanced disease and lower survival rates in minorities, have been described in various cancers, and colon cancer is not an exception to this [14]. Women generally fare better in terms of colon cancer incidence and outcomes [15]. Early detection is crucial in colon cancer, as 5 year survival can range from 14.5% to 97% depending on staging [16]. The introduction of universal healthcare in Massachusetts and increasing rates of screening colonoscopies nationwide, are factors that could induce major shifts in these differences. In fact, while evidence shows that screening reduces the overall incidence [17] and drastically improves the outcomes of colon cancer [18], differences have been reported in participation rates between whites and minorities [19,20], and men and women [21]. The objective of this study was to evaluate the state of disparities in health outcomes among colon cancer patients at Massachusetts General Hospital, a large, public hospital in a state that has phased in universal ß 2014 Wiley Periodicals, Inc. healthcare and where about a fifth of colon cancer diagnoses are made through screening. PATIENTS AND METHODS Study Setting Massachusetts General Hospital is located in a highly urbanized area with a diverse and multiethnic patient population. The Commonwealth of Massachusetts currently provides universal healthcare to its residents, and its MassHealth program has been granting even the lowest income groups access to care for over a decade: In 1997, under a Section 1115 waiver of the Social Security Act, MassHealth was Grant sponsor: The Fulbright Foundation, Harvard Catalyst The Harvard Clinical and Translational Science Center, the Dutch Cancer Society, the Dutch Digestive Society, and the Amsterdam University Funds. Conflicts of interest: None of the authors have conflicts of interest. Author contributions: DB had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. *Correspondence to: David L. Berger, MD, Division of General Surgery & Gastrointestinal Surgery, Massachusetts General Hospital, 15 Parkman Street, Boston, MA Fax: þ E mail: dberger@mgh. harvard.edu Received 04 December 2013; Accepted 31 December 2013 DOI /jso Published online in Wiley Online Library (wileyonlinelibrary.com).

2 2 Amri et al. introduced. This dramatically decreased the number of uninsured [22] in the state by providing insurance coverage to almost 1 million underprivileged residents (ca. 16% of the population) by 2004 [23]. When universal healthcare went through its last legislative step with the enactment of the Massachusetts Healthcare Insurance Reform Law of 2006, only 6% of the state was uninsured [23], and this number further decreased to 2.9% in the last Massachusetts Health Survey of that decade [24]. Study Design This is a prospectively maintained cohort study that includes all patients operated on for colon cancer with curative intent at our center from 2004 through Patients were identified through multiple data sources to ensure exhaustive inclusion. Sources included the MGH Cancer Registry and the Partners Healthcare Research Patient Data Registry (RPDR). All research followed a protocol approved by the MGH Institutional Review Board. The decision was made to include colon cancer patients only instead of all colorectal patients because colon and rectal cancer differ in staging, treatment protocols and stage specific outcomes [25]. The choice for 2004 as a cutoff date was made in order to take into account the expansion of MassHealth and a timeframe where the number of uninsured was already low and significantly below the national average, while still going back far enough to include over 1,000 patients. Data Collection All data was gathered through three full iterations of data collection in order to minimize entry errors and data distortion. All retrospectively accrued data was verified through multiple sources including records from the surgery, medical oncology, anesthesia, and pathology departments, as well as possible outside providers. Baseline characteristics collected included age, ethnicity, body mass index (BMI), gender, smoking status, and adjusted Charlson comorbidity score. The adjusted score used the definition by Charlson et al. [26], but excluded contributions of the patient s colon cancer to the score. This was done to circumvent the confounding effect of differences in metastatic disease, which greatly influence the score. In addition, type 2 diabetes (DM2) rates were also assessed as a separate comorbidity. We also assessed whether patients were diagnosed through screening, if they received preoperative chemotherapy, and whether admission was urgent. Following baseline presentation, perioperative characteristics including duration of stay and surgery, need for multivisceral resection during surgery (any additional visceral resection for oncologic reasons), disease progression, pathological characteristics, and need for adjuvant chemotherapy were assessed. Disease progression included clinically metastatic presentation, defined as metastatic disease established during preoperative imaging or surgical pathology. Other pathological characteristics measured were the AJCC TNM classification [27], tumor grade and microsatellite instability, rates of nodal disease, number of tumor positive lymph nodes, margin clearance and extent of resection. Long term outcomes were compared and defined as follows: Followup duration was the time from surgical intervention to the last visit that comprised a diagnostic intervention related to colon cancer follow up (usually serum CEA measurement, imaging and/or colonoscopy). Disease free survival was either the follow up duration or the interval between surgery and the first manifestations of recurrence or metastasis. Survival duration was the interval between the date of surgery and the last date a patient was known to be alive. Alive status and date of death relied on internal data and the Social Security Death Index. Deceased patients were subdivided into deaths directly attributable to the malignancy and those that died of other reasons. Statistical Analysis A P value of below 0.05 (two tailed) was considered significant. All statistical analysis was performed using the SPSS statistical suite (IBM Corp. Released IBM SPSS Statistics for Windows, Version Armonk, NY: IBM Corp.). A Mann Whitney U test was used to compare nonparametric ranked variables among nominal groups, such as the distribution of age or BMI over different ethnicities. The chisquare (x 2 ) statistic or the Fisher s exact test was used to compare these nominal groups with nominal and dichotomous variables (gender, smoking, comorbidities, etc.). The Cox proportional hazards model and logistic regression were used to more thoroughly assess for statistical significance of any differences encountered after correction for significant covariates. RESULTS Patient Demographics We included 1,071 patients, among those 521 (48.6%) females and 110 (10.3%) minority patients. Women had a lower median BMI (27.5 vs. 25.7; P < 0.001), were less likely to have a smoking history and had a lower comorbidity burden (mean adjusted Charlson score 0.71 vs. 0.80; P ¼ 0.021). However, women of screening age also had a significantly lower rate of screening diagnoses (26.4% vs. 32.7%; Relative risk: 0.80, 95% CI ; P ¼ 0.036). Minority groups were younger (median age 64 vs. 67 years; P ¼ 0.001), had a higher comorbid burden (mean adjusted Charlson score 0.83 vs. 0.72; P ¼ 0.039) largely owing to differences in comorbid DM2 rates (27.3% vs. 16.8%; P ¼ 0.006), and fewer former smokers (30.9% vs. 41.8%; P ¼ 0.027). Minorities also had a markedly higher number of patients that underwent neoadjuvant chemotherapy before admission (9.1 vs. 2.6; P ¼ 0.002). Baseline characteristics are shown in further detail in Table I. Perioperative Outcomes By ethnicity. As shown in Table II, minorities presented with significantly higher rates of lymph node metastasis (50% vs. 38.8%; P ¼ 0.014) related to far higher rates of N2 disease (25.5% vs. 14.8%; P ¼ 0.004), as illustrated by a higher median number of tumor positive lymph nodes (4 vs. 3; P ¼ 0.06). Minorities also had higher rates of clinically metastatic presentation (21.8% vs. 14%; P ¼ 0.026), with longer mean durations of surgery (168 vs. 140 min; P ¼ 0.014) and a longer surgical stay (7 vs. 6.6 days; P ¼ 0.019). A survey of internal complication data did not show any statistically significant differences in complication rates that could explain this through a higher rate of postoperative complications. By gender. Women had significantly higher rates of M stage 1 disease on pathology (11.3% vs. 7.1%; P ¼ 0.02) and had noteworthy differences in rates of clinically metastatic presentation (16.1% vs. 13.7%; P ¼ 0.27) high grade tumors on pathology (22.1% vs. 17.1%; P ¼ 0.08), microsatellite instability (45.2% vs. 29.1%; P ¼ 0.08) and rates of radical resection (88.1% vs. 91.2%; P ¼ 0.09). Women were also noted to have significantly faster surgeries (median 135 vs. 151 min; P < 0.001) despite a higher rate of multivisceral resections (17.4% vs. 10.4%; P < 0.001). Long Term Outcomes Absolute numbers show small yet consistently unfavorable differences for women across all outcomes. Substantial disparities detrimental to minorities were found in median survival (142.5 vs. 185 weeks; P ¼ 0.022) and disease free survival (57.5 vs. 89 weeks; P ¼ 0.013). Differences in overall metastatic disease were borderline significant (33.6% vs. 25.5%; P ¼ 0.055), and come along with a notable difference in terms of cancer related mortality (21.8% vs. 17.5%;

3 Colon Cancer Disparities & Care Access 3 TABLE I. Baseline Characteristics at Presentation White Minority P value Male Female P value Age (years; M, IQR) BMI (kg/m 2 ; M, IQR) <0.001 Gender (male; %) n/a n/a n/a Current smoking (%) Former smoker (%) <0.001 Adjusted Charlson score (x, SD) Comorbid type 2 diabetes (%) Screening diagnosis all (%) 20.5 (<:29.8,:24.6) 18.2 (<: 31.5,: 16.1) Screening diagnosis primary (%) 29.9 (<: 32.5,: 27.3) 26.6 (<: 34.8,: 18.8) Urgent admission (%) Neoadjuvant chemotherapy (%) F M, median; IQR, interquartile range; x, mean; SD, standard deviation; F : Fischer s exact test. Primary target group for screening: patients 50, no prior CRC history; N ¼ 919. </,: screening rates for male and females separately in this subset. P ¼ 0.25), despite nearly similar post presentation metastasis rates (11.8% vs. 11.4%, RR ¼ 1.01; P ¼ 0.91) and overall mortality rates (34.5% vs. 32.2%, RR ¼ 1.07; P ¼ 0.49). Kaplan Meier survival estimates as shown on Figure 1 echo the small differences in median survival, especially in the long term: At 3 years, survival estimates were 69.7% in minorities versus in 75.9% whites; 5 year survival estimates were 63% versus 66.3%. All results with respective relative risks and confidence intervals are shown in Table III. Multivariate Analysis Multivariate models in Table IV comparing genders yielded no significant results, although point estimates and confidence intervals were unanimously skewed towards less favorable outcomes for women that are mitigated by adjustment for metastatic presentation and AJCC classification. For white minority differences, point estimates are reduced considerably and to very small values upon adjustment for metastatic presentation, and even more so after AJCC staging, displaying the cardinal role of differences at presentation in the outcome differences shown: Metastatic disease estimate decreases from (OR ¼ 1.52, 95% CI ; P ¼ 0.051) to OR ¼ 1.13 (95% CI ; P ¼ 0.67); overall mortality from OR ¼ (95% CI ; P ¼ 0.22) to OR ¼ 1.10 (95% CI ; P ¼ 0.72) corrected for age, Charlson comorbidity score and date of surgery; and cancer related mortality from OR ¼ 1.31 (95% CI ; P ¼ 0.26) to OR ¼ 1; (95% CI ; P ¼ 0.98). Cox regression shows similar decreases over time dependent risk of death and metastasis. Follow up (FU) duration, a potential confounder for shorter survival and disease free survival as it is significantly shorter in minorities (B ¼ 182 days, P ¼ 0.02) was no longer significantly different after adjusting for date of surgery and metastatic presentation (B ¼ 69 days. 95% CI: 200, 60; P ¼ 0.29), confirming that differences in FU are chiefly due to differences in metastatic presentation along with a higher influx of minority patients in recent years and confirming that date of surgery is a better covariate choice than follow up duration. Cox proportional hazards models did not need adjustment for this effect at all, as they are standardized for units of time [28]. Lastly, the Kaplan Meier survival curve (Figure 1) illustrates why stage for stage survival over time demonstrates the absence of a persistent discrepancy pattern, with minorities performing slightly better in stage I (P ¼ 0.15) and IV (P ¼ 0.57), and slightly worse in stage II (P ¼ 0.28) and III (P ¼ 0.23). DISCUSSION We aimed to measure ethnic or gender disparities that may exist in patients surgically treated for colon cancer at Massachusetts General TABLE II. Perioperative Outcomes White Minority RR (95% CI) P value Men Women RR (95% CI) P value Transmural spread (% T4) ( ) ( ) 0.28 Positive LN a (%) ( ) ( ) Positive LN b (N, M, IQR) n/a n/a 0.71 N1 disease (%) ( ) ( ) 0.47 N2 disease (%) ( ) ( ) 0.73 M1 disease (%) ( ) ( ) 0.02 Radical (R0) resection (%) (0.67 2) ( ) 0.09 Clearance (cm, M, IQR) n/a n/a 0.78 High grade (%) ( ) ( ) 0.08 MSI positive (%, n tested) 40 (110) 0 (7) 0.15 ( ) (55) 45.2 (62) 1.55 ( ) 0.08 Metastatic presentation c (%) ( ) ( ) 0.27 Surgery duration (min; x, SD) n/a n/a Multivisceral resection (%) ( ) ( ) Stay duration (days; x, SD) n/a n/a 0.54 Complicated stay (%) ( ) ( ) Postoperative chemotherapy (%) ( ) ( ) 0.75 M, median; IQR, interquartile range; x, mean; SD, standard deviation; RR, relative risk; MSI, microsatellite instability; LN, Lymph node. a Excludes stage N1c tumors. b Number of tumor positive nodes in LNþ resections. c Clinically metastatic, includes Mx resections.

4 4 Amri et al. Fig. 1. Kaplan Meier survival curve subdividing both white and nonwhite populations in AJCC staging subgroups in order to compare these. Overall, minorities with stage I and stage IV disease fare slightly better, while the opposite is true for stage II and III. More importantly, the differences are slight and none of the comparisons show significant differences, as opposed to overall survival statistics. The embedded table show year by year survival estimates for all four stages and overall for minorities and white patients. TABLE III. Comparison of Long Term Outcomes White Minority P value Men Women P value Disease free survival, weeks (M) Survival duration, weeks (M) White Minority RR (95% CI) P value Men Women RR (95% CI) P value Metastasis after presentation (%) ( ) ( ) 0.69 Overall metastasis (%) ( ) ( ) 0.55 Overall mortality (%) ( ) ( ) 0.36 Cancer related mortality (%) ( ) ( ) 0.40 M, median; IQR, interquartile range; x, mean; SD, standard deviation. OR, odds ratio; RR, relative risk.

5 Colon Cancer Disparities & Care Access 5 TABLE IV. Univariate and Multivariate Regression Analysis of Long Term Outcomes Non white vs. White Female vs. Male Cox regression Covariates HR (95% CI) P Covariates HR (95% CI) P Disease free survival None 1.43 ( ) None 1.09 ( ) 0.48 Mþ Presentation 1.11 ( ) 0.54 Mþ Presentation 1.04 ( ) 0.94 AJCC 1.07 ( ) 0.72 AJCC 0.99 ( ) 0.95 Survival None 1.20 ( ) None 1.10 ( ) 0.34 Age, Charlson 1.23 ( ) 0.23 Charlson 1.14 ( ) 0.23 Age, Charlson,MþP 0.99 ( ) 0.95 Charlson,Mþ P 0.94 ( ) 0.62 Linear regression Covariates B (95% CI) P Covariates OR (95% CI) P Follow up duration None 182 ( 336, 27) 0.02 None 48 ( 45, 141) 0.31 Date of surgery, 102 ( 235, 31) 0.13 Charlson,Mþ P 46 ( 44,136) 0.31 DoS,Mþ Presentation 69 ( 200, 60) 0.29 Logistic regression Covariates OR (95% CI) P Covariates OR (95% CI) P Metastatic disease Date of surgery 1.52 ( ) None 1.09 ( ) 0.55 DoS,Mþ presentation 1.20 ( ) 0.56 DoS, AJCC 1.13 ( ) 0.67 Overall mortality Age, Charlson, DoS 1.32 ( ) 0.22 None 1.12 ( ) 0.36 idem þ AJCC 1.10 ( ) 0.72 Charlson 1.15 ( ) 0.27 Cancer related mortality None 1.31 ( ) 0.26 None 1.06 ( ) 0.73 Mþ Presentation 0.99 ( ) 0.97 Met. Presentation 1.05 ( ) 0.82 DoS, AJCC, 1 ( ) 0.99 Significance of covariate in model: P < 0.05; P < B, unstandardized regression coefficient, difference estimate in minutes; AJCC, American Joint Committee on Cancer staging; Charlson, Charlson Comorbidity Score (adjusted for colon cancer); DoS, date of surgery; Mþ P/M þ Presentation, presentation with metastatic disease. Hospital. The baseline characteristics of our population follow agematched minority distributions in Massachusetts in recent census data, where about 9% of people over 65 in the state were found to belong to a non white ethnic group [29]. We found that women were diagnosed significantly less often through screening means, and more often urgently, both predictors of worse outcomes [18,30]. Perioperatively, minorities, who came in with more comorbid disease and especially higher DM2 numbers had higher rates of nodal disease and metastasis, along with longer surgeries and longer surgical stays, which were not explained by either higher rates of multivisceral resections or postoperative complications. Women had higher rates of metastatic disease in pathology, as well as higher percentages of high grade and microsatellite unstable tumors. In terms of long term outcomes, minorities had higher overall rates of metastatic disease, death and cancer related deaths, as well as shorter disease free survival and survival duration. In sum, disparities at presentation as well as long term outcomes were omnipresent. Once disparities at presentation were accounted for, the disparities in outcome disappeared; stage for stage treatment and follow up rates were not substantially different based upon ethnicity or gender supporting the hypothesis that differences in outcomes may be largely due to factors like late presentation and underrepresentation in screening diagnosis. Consequences of the Findings Significant ethnic and gender inequalities remain at presentation after more than a decade of tremendous efforts to achieve an equal footing in healthcare in Massachusetts. Differences in long term outcomes, although less flagrant, were still visible. A reassuring finding however, is that disparities are mitigated once patients were enrolled into care. A nonrestrictive healthcare system, with expertise inherent to a tertiary care setting made it possible to achieve comparable stagespecific treatment outcomes despite an otherwise unequal footing at presentation. These findings are also relevant to the United States at large, who may soon face healthcare changes with many similarities to the policy changes in our state. Limitations Statistical power is the unsurprising core limitation of this study. When facing equivalence in outcomes, several findings were not statistically significant, especially since common statistical methods are designed to prove significance of differences. Minority to white patient comparisons were most affected because of their proportions. Often, the numbers themselves did carry a value beyond the judgment of a P value. Aside from the intrinsic value of many of the percentage outcomes, point estimates and 95% confidence intervals that consistently stay within certain ranges are also informative. In addition, possible motives to rely heavily on P values for outcome interpretation are less relevant to this type of research. There is no selection bias, as this was a complete crosssection of all surgical colon cancer patients in the given timeframe. There is no substantial risk of interpretation bias as outcomes and staging characteristics are based on objective data outside the control of the authors. Lastly, with over 1,000 patients in an 8 year timeframe, this is probably one of the largest samples achievable without inter center variability or time related changes in treatment approach or population makeup. CONCLUSIONS The outcomes described in this paper show that equality in outcome of provided care is achievable despite unequal footing at presentation. Nationwide statistics describe serious discrepancies that continue to worsen after enrollment into care [31 34]; Individual institutions should not consider this a fait accompli. In fact, universal health care and

6 6 Amri et al. sustained access to tertiary care may level the playing field sufficiently to keep outcome gaps at near undetectable levels, even in a large population sample. Cancer treatment is tremendously expensive, and lack of access through lack of insurance coverage is directly associated with worse cancer staging and survival [35]. Since close to 50 million people in the US are uninsured [36], a great many of these are receiving suboptimal care, resulting in disparities in both their presentation and their outcomes. The combination of being a public city hospital with a commitment to serve the community and considerable effort by the State to provide universal coverage were able to provide the low threshold care necessary to mitigate inequality at presentation; however, most of the gains, and the crux of providing healthcare to all will be in prevention and early detection. Over time, the impact of universal access may for instance improve screening rates in minorities and therefore have a greater effect on reducing discrepancies on presentation with colon cancer. Meanwhile, as our example also illustrates, we also need to remain realistic and understand that universal healthcare will not be an instant solution to the deeply rooted issues that cause inequalities at presentation. These should not be dismissed by the argument that (more) universal healthcare is on its way, nor should universal healthcare be seen as the panacea for these issues. Many of the discrepancies faced in caring for a heterogeneous population come from origins far more complex than access to care, and the root issues often lie deeper than financing. Healthcare practitioners should also remain alert for shifts during significant changes to the system, as these may lead to new inequalities, warranting reporting or requiring action. ACKNOWLEDGMENTS This work was conducted with support from Harvard Catalyst, The Harvard Clinical and Translational Science Center (National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health Award 8UL1TR and financial contributions from Harvard University and its affiliated academic health care centers). The content is solely the responsibility of the authors and does not necessarily represent the official views of Harvard Catalyst, Harvard University and its affiliated academic health care centers, or the National Institutes of Health. REFERENCES 1. Geiger HJ, Borchelt G: Racial and ethnic disparities in US health care. Lancet 2003;362:1674. doi: /S (03) Dunlop DD, Manheim LM, Song J, et al.: Gender and ethnic/racial disparities in health care utilization among older adults. J Gerontol B Psychol Sci Soc Sci 2002;57:S221 S Read JG, Gorman BK: Gender inequalities in US adult health: The interplay of race and ethnicity. Soc Sci Med 2006;62: doi: /j.socscimed Wong MD, Shapiro MF, Boscardin WJ, et al.: Contribution of major diseases to disparities in mortality. N Engl J Med 2002;347: doi: /NEJMsa Paasche Orlow MK, Parker RM, Gazmararian JA, et al.: The prevalence of limited health literacy. J Gen Intern Med 2005;20: doi: /j x 6. Bach PB, Pham HH, Schrag D, et al.: Primary care physicians who treat blacks and whites. N Engl J Med 2004;351: doi: /NEJMsa Siegel S, Moy E, Burstin H: Assessing the nation s progress toward elimination of disparities in health care. J Gen Intern Med 2004; 19: Harper S, Rushani D, Kaufman JS: Trends in the black white life expectancy gap, JAMA 2012;307: doi: /jama Shellnut JK, Wasvary HJ, Grodsky MB, et al.: Evaluating the age distribution of patients with colorectal cancer: Are the United States Preventative Services Task Force guidelines for colorectal cancer screening appropriate? Dis Colon Rectum 2010;53:5 8. doi: /DCR.0b013e3181bc01d1 10. Haggar FA, Boushey RP: Colorectal cancer epidemiology: Incidence, mortality, survival, and risk factors. Clin Colon Rectal Surg 2009;22: doi: /s Giovannucci E, Harlan DM, Archer MC, et al.: Diabetes and cancer: a consensus report. Diabetes Care 2010;33: doi: /dc Kushi LHL, Byers TT, Doyle CC, et al.: American Cancer Society Guidelines on Nutrition and Physical Activity for cancer prevention: Reducing the risk of cancer with healthy food choices and physical activity. CA Cancer J Clin 2006;56: doi: / canjclin Siegel R, Naishadham D, Jemal A: Cancer statistics, CA Cancer J Clin 2012;62: doi: /caac Ward EE, Jemal AA, Cokkinides VV, et al.: Cancer disparities by race/ethnicity and socioeconomic status. CA Cancer J Clin 2004;54: doi: /canjclin Murphy G, Devesa SS, Cross AJ, et al.: Sex disparities in colorectal cancer incidence by anatomic subsite, race and age. Int J Cancer 2011;128: doi: /ijc Gunderson LL, Jessup JM, Sargent DJ, et al.: Revised TN categorization for colon cancer based on national survival outcomes data. J Clin Oncol 2010;28: doi: / JCO Siegel RL, Ward EM, Jemal A: Trends in colorectal cancer incidence rates in the United States by tumor location and stage, Cancer Epidemiol Biomarkers Prev 2012;21: doi: / EPI Amri R, Bordeianou LG, Sylla P, et al.: Impact of screening colonoscopy on outcomes in colon cancer surgery. JAMA Surg 2013;148: doi: /jamasurg Laiyemo AO, Doubeni C, Pinsky PF, et al.: Race and colorectal cancer disparities: Health care utilization vs different cancer susceptibilities. J Natl Cancer Inst 2010;102: doi: /jnci/djq Breen NN, Wagener DKD, Brown MLM, et al.: Progress in cancer screening over a decade: Results of cancer screening from the 1987, 1992 and 1998 National Health Interview Surveys. J Natl Cancer Inst 2001;93: Peterson NB, Murff HJ, Ness RM, et al.: Colorectal cancer screening among men and women in the United States. J Womens Health (Larchmt) 2007;16: doi: /jwh Long SK, Zuckerman S: MassHealth succeeds in expanding coverage for adults. Inquiry 2004;41: Seifert RW, Anthony S: The Basics of MassHealth. Massachusetts Medicaid Policy Institute 2011; Long SK, Stockley K, Dahlen H: Massachusetts health reforms: Uninsurance remains low, self reported health status improves as state prepares to tackle costs. Health Aff (Millwood) 2012;31: doi: /hlthaff Li M, Li JY, Zhao AL, et al.: Colorectal cancer or colon and rectal cancer? Clinicopathological comparison between colonic and rectal carcinomas. Oncology 2007;73: doi: / Charlson ME, Pompei P, Ales KL, et al.: A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 1987;40: Edge SB, Byrd DR, Compton CC, et al.: AJCC cancer staging manual. 7th edition. New York, NY: Springer; Bewick V, Cheek L, Ball J: Statistics review 12: Survival analysis. Crit Care 2004;8: doi: /cc U.S. Bureau of the Census American community survey 1 year estimates. Detailed tables generated by Ramzi Amri using American Factfinder; November 20, : Elliss Brookes L, McPhail S, Ives A, et al.: Routes to diagnosis for cancer Determining the patient journey using multiple routine data sets. Br J Cancer 2012;107: doi: / bjc

7 Colon Cancer Disparities & Care Access Marcella S, Miller JE: Racial differences in colorectal cancer mortality. The importance of stage and socioeconomic status. J Clin Epidemiol 2001;54: Alexander DD, Waterbor J, Hughes T, et al.: African American and Caucasian disparities in colorectal cancer mortality and survival by data source: An epidemiologic review. Cancer Biomark 2007;3: Roetzheim RG, Pal N, Gonzalez EC, et al.: Effects of health insurance and race on colorectal cancer treatments and outcomes. Am J Public Health 2000;90: Mayberry RM, Coates RJ, Hill HA, et al.: Determinants of black/ white differences in colon cancer survival. J Natl Cancer Inst 1995;87: Halpern MT, Ward EM, Pavluck AL, et al.: Association of insurance status and ethnicity with cancer stage at diagnosis for 12 cancer sites: A retrospective analysis. Lancet Oncol 2008;9: doi: /S (08) U. S. Bureau of the Census. Income, poverty, and health insurance coverage in the United States: US Government Printing Office. 2011;1 95.