Exploring the role of marital status on survival and stage at diagnosis in early-onset colorectal cancer by race/ethnicity

Abstract

We examined the relationship of marital status on cause-specific survival and late-stage at diagnosis for early-onset colorecta cancer (EO-CRC), overall and within five racial/ethnic groups (non-Hispanic White [NHW], non-Hispanic Black [NHB], American Indian/Alaskan Native [AI/AN], Asian/ Pacific Islanders [PI], or Hispanic). We conducted a retrospective cohort analysis using data from 2006 to 2015 Surveillance, Epidemiology, and End Results Program. The Kaplan-Meier method, multivariable Cox proportional hazards regression, and logistical regression models were performed. Among 58,884 participants, the lower five-year survival rate was 66.4% for EO-CRC patients who were unmarried (p-value < 0.001). The 5-year survival rate was the lowest for NHB patients (64.9% vs. 69.7% for AI/AN, 70.1% for Hispanic, 72.2% for Asian/PI, and 72.8% for NHW). Unmarried NHW (hazard ratio [HR], 1.25; 95% CI, 1.19–1.30), NHB (HR, 1.28; 95% CI, 1.18–1.39), Asian/PI (HR, 1.17; 95% CI, 1.03–1.32), and Hispanic patients (HR, 1.10; 95% CI, 1.01–1.19) were more likely to die from CRC than those were married. Further, unmarried NHW (OR, 1.16; 95% CI, 1.09–1.22), NHB (OR, 1.23; 95% CI, 1.10–1.38), and Asian/PI patients (OR, 1.18; 95% CI, 1.00-1.38) had a risk for a later stage at diagnosis for CRC. We observed lowest five-year survival rate among unmarried patients. Culturally sensitive interventions aimed at improving EO-CRC survival may benefit individuals across all racial groups. Such efforts can be strengthened by enhancing social support networks and increasing CRC risk awareness as well as equitable access to care, particularly for NHW, NHB, and Asian/PI populations.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-51116-7.

Introduction

Early-onset colorectal cancer (EO-CRC, defined as diagnosed with CRC at age < 50 years) has increased significantly in the United States (US)¹. The incidence of CRC is predicted to increase by 90% by 2030 in individuals aged 20–34 years². This has been associated with an increase in EO-CRC mortality by approximately 1.3% per year from 2008 to 2017³. Further, racial/ethnic disparities in EO-CRC survival and later stage at diagnosis were also observed, with significantly worse survival and greater later stage at diagnosis among Non-Hispanic Blacks (NHBs)^4,5. Although the factors associated with EO-CRC diagnosis and mortality are still unclear, contributing factors may include sociodemographic characteristics^6–9, lifestyle factors (e.g., physical activity, dietary pattern, smoking behaviors, alcohol use)^6,9–13, comorbidities (e.g., obesity, diabetes)^1,7,14, and family history of CRC¹⁵.

Marital status has been linked to differences in social support that may influence psychosocial well-being and cancer outcomes, including the possibility of selection bias^16–19. Married patients tend to have better survival partly because they are more likely to receive earlier diagnoses and appropriate treatment, while unmarried patients face higher risks of delayed diagnosis and poorer outcomes¹⁸. Additional evidence showed that unmarried patients were more likely to have a delayed diagnosis and experience faster cancer progression than married counterparts¹⁶. Several international studies also confirmed this association for risk of CRC death^16,20–22. Recently, a Utah study examined the social determinants of EO-CRC survival and suggested the important role of marital status for social support among young populations²³. Because NHB patients were associated with 35% increased risk of cause specific mortality for EO-CRC⁴, research examining the association of marital status on EO-CRC outcomes when considering different racial/ethnic differences is necessary.

Despite the existence of evidence, prior studies have primarily focused on patients with a wide age range including EO-CRC and CRC diagnosed at or after 50 years of age or on non- US populations^16,23,24. Limited research has also examined EO-CRC survival disparities in other minorities (e.g., Asian or Hispanic groups)^4,25,26. Only one US study examined the role of marital status on EO-CRC survival across five racial/ethnic groups, limited to adenocarcinoma cases²⁷. To address these gaps, we sought to examine the relationship of marital status on cause-specific survival and late stage at diagnosis for EO-CRC, overall and across five racial/ethnic groups (NHW, NHB, American Indian/Alaskan Native [AI/AN], Asian/ Pacific Islanders [PI], or Hispanic). We also compared the 5-year cause-specific survival rates for CRC across these five racial/ethnic groups and considered differences in marital status. Given the important role that marital status as a measure of social support, such understanding is crucial for guiding successful culturally sensitive interventions aimed at improving CRC risk awareness and promoting equitable access to health care for young populations through their support networks.

Methods

Study design

We conducted a retrospective cohort analysis using data from the 2006–2020 Incidence Data with Census Tract Attributes from Surveillance, Epidemiology, and End Results (SEER) Program (22 registries), which are sources for comprehensive population-based information in the U.S. that includes patient demographics, census tract level persistent poverty and rurality, primary tumor site, tumor morphology and stage at diagnosis, first course of treatment, and follow-up for vital status. The eligible study population included patients diagnosed with CRC defined by the SEER Site Recode ICD-O-3/WHO 2008 definition of colon cancer (C180–C189), rectosigmoid junction cancer (C199), and rectal cancer (C209)²⁸. Data extracted for this study were publicly available and de-identified, and thus considered exempt from institutional review board (IRB) review at Augusta University.

Study participants

A total of 835,907 CRC patients were included in the 2006–2020 SEER program. To obtain the eligible study sample for EO-CRC patients, we first excluded patients who were diagnosed at age of ≥ 50 years (n = 482,098). We then excluded duplicate records for the same patients with same diagnosis (n = 11,338), those with missing information on rurality (n = 22) and persistent poverty (n = 248), CRC diagnosed after 2015 due to limited follow-up time (less than five years) (n = 275,291), missing survival time (months) from SEER program (n = 5,913), missing vital status (n = 18), and missing race/ethnicity (n = 2,095). As a result, 58,884 eligible EO-CRC patients diagnosed with CRC were included in full analyses.

Outcomes, exposures, and covariates

Cause-specific survival and late stage at diagnosis for CRC were the outcomes of interest. Cause-specific death for CRC was defined by using SEER cause-specific death classification. SEER staging was used to define localized, regionalized, and distant diagnoses for stage at diagnosis of CRC. Both regionalized and distant stages of CRC were classified as late stages of diagnosis (termed as yes) and the rest of diagnosis as no. Further, race/ethnicity and marital status were two exposures of interest. Patients’ race/ethnicity was defined as NHW, NHB, AI/AN, Asian/PI, or Hispanic. SEER codes race and Hispanic ethnicity as separate variables; therefore, we used a combined race/ethnicity measure with categories of NHW, NHB, AI/AN, Asian/PI, or Hispanic (any race). We classified marital status as a three-level variable, including married (married/partnered), unmarried (single/widowed/divorced), and unknown.

Covariates of interest included demographic characteristics, cancer-related factors, and tumor features. For demographic characteristics, we included age at diagnosis (20–29, 30–39, 40–49 years), sex (male or female), census tract persistent poverty (yes or no), census tract rurality (yes or no). The definition of persistent poverty was based on 1990, 2000 decennial censuses, and 2007-11 and 2015-19 American Community Survey 5-year estimates. Based on this, persistent poverty was defined as 20% or more of the population living below the poverty level for a period spanning about 30 years. Areas meeting this criterion were classified as “persistent poverty (termed as yes),” while those that did not were classified as “non-persistent poverty (termed as no). Rurality was defined by using the U.S. Department of Agriculture (USDA)’s Rural Urban Commuting Area (RUCA) codes, with 1.0, 1.1, 2.0, 2.1, 3.0, 4.1, 5.1, 7.1, 8.1, and 10.1 were considered as urban areas (termed as no) and all other codes were classified as rural areas (termed as yes). For cancer related factors, we included years of diagnosis (before 2011 or 2011 or later) and first malignant primary cancer (yes or no). For tumor features, we included pathological grading (well differentiated, moderately differentiated, poorly differentiated, undifferentiated, or unknown) and primary site (left or right). All covariates of interest—including demographic characteristics, cancer-related factors, and tumor features—were used to adjust for potential confounding rather than to compare case characteristics, and are presented in Supplementary Table 1.

Multiple imputation

After the exclusion of sample based on the above criteria, marital status (our exposure of interest) was variable with a high proportion of unknown data. Multiple imputation by fully conditional specification (FCS MI) was used to approximate marital status for unknown values (n = 14,045, 23.9%)^16,29. FCS MI is a useful and statistically valid approach for creating imputations in large epidemiologic studies for which data include both categorical and continuous variables³⁰. Further, the SAS multiple imputation procedure was used to generate 15 imputed data for missing marital status data with Cox proportional hazard regression and logistic regression imputation methods. By applying multiple imputation, potential bias can be reduced; simulation studies comparing multiple imputation to complete case analyses suggest that excluding records with missing data can considerably bias regression coefficients³¹.

Statistical analysis

Descriptive statistics were used to describe the distribution of EO-CRC patients according to race/ethnicity, marital status, demographic characteristics, cancer-related factors, and tumor features. Bivariate differences were also examined using chi-square tests. Patients’ survival time was measured in months from the month of diagnosis up to 60-months of follow up or until death due to CRC, censored at the end of the study observation period (December 31, 2020) or death due to other causes. Survival analysis at five-year intervals was performed using the Kaplan-Meier method. The Log-rank test was used to compare the survival rates within five racial/ethnic groups considering differences of marital status based on MI results. Further, we performed Cox proportional hazard regression and logistic regression to examine the association between race/ethnicity and marital status on cause-specific survival and late stage at diagnosis for CRC, respectively. A complete-case model excluding records with unknown marital status and MI model were developed to examine the mentioned association. Finally, subpopulation analyses within five racial/ethnic groups for the association between marital status, risk of CRC death, and late stage at diagnosis were also performed using MI models. All multivariable models were adjusted for demographic characteristics, including age at diagnosis (using SEER’s categorical age groups), as well as cancer-related factors and tumor features, to minimize potential confounding. Hazard ratio (HR) and odd ratio (OR) associated with 95% confidence interval (CI) were obtained. The level of statistical significance was set at an alpha level of 0.05 and the p-values were based on two-sided probability tests. We used SAS version 9.4, SAS Institute Inc., Cary, North Carolina, and IBM SPSS version 28.0 to perform all the analyses.

Results

Characteristics of study population

As shown in Table 1 and Supplementary Table 1, most EO-CRC patients were married (44.6%), aged 40–49 years (73.3%), male (52.5%), lived in non-persistent poverty (88.8%), and non-rural (88.8%) areas. Many were diagnosed with CRC 2011or later (51.6%), had first malignant primary cancer (94.1%), had a regionalized diagnosis (37.4%), had a moderately differentiated diagnosis (37.4%), and had a left-sided CRC (71.2%). When examining racial/ethnic differences, a high proportion of married patients were Asian/PI (60.4%), with lower proportions of married patients among NHB (44.9%) and AI/AN groups (44.4%) (p-value < 0.001).

Table 1.

Characteristics of study population by race/ethnicity.

	Total (n = 58,884)	NHW (n = 33,557, 57%)	NHB (n = 8,469, 14.4%)	Hispanic (n = 11,604, 19.7%)	Asian/PI (n = 4,891, 8.3%)	AI/AN (n = 363, 0.6%)	P-value
	n (%)
Marital status							< 0.001
Married	26,288(44.6%)	16,409(48.9%)	2,514(29.7%)	4,278(36.9%)	2,955(60.4%)	132(36.4%)
Unmarried a	18,551(31.5%)	9,917(30.0%)	3,804(44.9%)	3,369(29.0%)	1,300(26.6%)	161(44.4%)
Unknown	14,045(23.9%)	7,231(21.6%)	2,151(25.4%)	3,957(34.1%)	636(13.0%)	70(19.3%)

CRC, colorectal cancer; NHW, non-Hispanic White; NHB, non-Hispanic Black; AI/AN, American Indian/Alaska Native; PI, pacific islander.

^a Unmarried includes single, divorced, separated, and widow.

Five-year survival rates

The mean survival time since EO-CRC diagnosis for married patients was 50.1 months and unmarried patients was 46.7 months (p-value < 0.001, overall: 49.1 months). As shown in Supplementary Table 2, the overall cause-specific survival rate at 5-year was 71.0% for the EO-CRC patients, with lower survival for unmarried patients (66.4%) (vs. married patients, p-value < 0.001). By racial/ethnic groups, the 5-year survival rate was the lowest for NHB patients (64.9%), followed by 69.7% for AI/AN, 70.1% for Hispanic, 72.2% for Asian/PI, and 72.8% for NHW patients. When exploring the survival rates according to marital status, the five-year survival rate was the lowest for unmarried patients at 68.5% (vs. 74.6% for married patients) among NHW patients (Fig. 1a, p-value < 0.001). Similarly, among NHB patients, the unmarried patients had the lowest survival rates (60.3% vs. 68.6% for married patients) (Fig. 1b, p-value < 0.001). AI/AN (62.7% vs. 75.2%, Fig. 1c, p-value = 0.007) and Asian/PI patients (67% vs. 74.1%, Fig. 1d, p-value < 0.001) who were unmarried also had lower 5-year survival rates compared to those who were married. Similarly, the lower survival rates were observed for Hispanic patients who were unmarried (67.4% vs. 71.2% for married patients) (Fig. 1e, p-value < 0.001).

Fig. 1.

Kaplan Meier colorectal cancer survival curves by marital status and race/ethnicity (n = 58,884). Notes: Given that the majority of EO-CRC patients were married based on the imputation results, patients were married at diagnosis was used to replace those with unknown marital status.

Multivariable analysis

In Table 2, results from complete case and MI models were highly consistent for both risk of CRC death and late stage at diagnosis, indicating that the imputed estimates closely aligned with the non-imputed results. In MI models, NHB patients were 1.4-fold (HR, 1.36; 95% CI, 1.30–1.42), AI/AN patients were 1.1-fold (HR, 1.08; 95% CI, 0.90–1.31), and Hispanic patients were 1.1-fold (HR, 1.12; 95% CI, 1.07–1.16) more likely to die from CRC compared to NHW patients. Unmarried patients were associated with increased risk of CRC death by 1.2 times than married patients (HR, 1.21; 95% CI, 1.17–1.26). Similarly, NHB (OR, 1.07; 95% CI, 1.01–1.13) and Hispanic (OR, 1.12; 95% CI, 1.07–1.18) patients were 1.1-fold more likely to have late stage at diagnosis than the NHW patients. Unmarried patients were also associated with a greater likelihood of having a late stage at diagnosis than married patients by 14% (OR, 1.14; 95% CI, 1.10–1.18). An additional analysis separating individuals who were single from those who were divorced, separated, or widowed showed consistent results for both risk of CRC-specific death and likelihood of late-stage diagnosis (Supplementary Table 3).

Table 2.

Association between race/ethnicity and marital status on the risk of CRC death and late stage at diagnosis.

	Risk of CRC death		Late stage at diagnosis
	Complete case model a (n = 44,839)	Multiple imputation model a (n = 58,884)	Complete case model a (n = 44,839)	Multiple imputation model a (n = 58,884)
	HR (95%CI)	HR (95%CI)	OR (95%CI)	OR (95%CI)
Race/ethnicity
NHW	Reference	Reference	Reference	Reference
NHB	1.36(1.30,1.43)	1.36(1.30,1.42)	1.05(0.99,1.12)	1.07(1.01,1.13)
Hispanic	1.17(1.11,1.23)	1.12(1.07,1.16)	1.10(1.04,1.17)	1.12(1.07,1.18)
Asian/PI	1.10(1.04,1.17)	1.06(1.00,1.13)	1.08(1.00,1.16)	1.06(0.99,1.14)
AI/AN	1.07(0.87,1.32)	1.08(0.90,1.31)	1.06(0.81,1.38)	0.94(0.74,1.19)
Marital status
Married	Reference	Reference	Reference	Reference
Unmarried	1.29(1.24,1.34)	1.21(1.17,1.26)	1.20(1.15,1.25)	1.14(1.10,1.19)

NA, non-applicable; CRC, colorectal cancer; HR, hazard ratio; OR, odd ratio; NHW, non-Hispanic White; NHB, non-Hispanic Black; AI/AN, American Indian/Alaska Native; PI, Pacific Islander.

Italicized text indicates statistically significant result.

^a All models were adjusted for demographic characteristics, cancer related factors, and tumor features.

As shown in Table 3, unmarried NHW (HR, 1.25; 95% CI, 1.19–1.30), NHB (HR, 1.28; 95% CI, 1.18–1.39), Asian/PI (HR, 1.17; 95% CI, 1.03–1.32), and Hispanic patients (HR, 1.10; 95% CI, 1.01–1.19) were more likely to die from CRC than those who were married. Further, unmarried NHW (OR, 1.16; 95% CI, 1.09–1.22), NHB (OR, 1.23; 95% CI, 1.10–1.38), and Asian/PI patients (OR, 1.18; 95% CI, 1.00-1.38) were associated with greater likelihood of having late stage at diagnosis for CRC than those who were married. An additional analysis examining sex differences showed that both male and female patients were more likely to die from CRC and to be diagnosed at a late stage (Supplementary Table 4).

Table 3.

Association between race/ethnicity and marital status on the risk of CRC death (n = 58,884).

	NHW	NHB	Hispanic	Asian/PI	AI/AN
Risk of death a
	HR (95% CI)
Marital status
Married	Reference	Reference	Reference	Reference	Reference
Unmarried	1.25(1.19,1.30)	1.28(1.18,1.39)	1.10(1.01,1.19)	1.17(1.03,1.32)	1.20(0.79,1.82)
Late stage at diagnosis a
	OR (95% CI)
Marital status
Married	Reference	Reference	Reference	Reference	Reference
Unmarried	1.16(1.09,1.22)	1.23(1.10,1.38)	1.09(0.99,1.22)	1.18(1.00,1.38)	1.14(0.63,2.07)

CRC, colorectal cancer; HR, hazard ratio; OR, odd ratio; NHW, non-Hispanic White; NHB, non-Hispanic Black; AI/AN, American Indian/Alaska Native; PI, Pacific Islander.

Italicized text indicates statistically significant results.

^a All models were adjusted for demographic characteristics, cancer related factors, and tumor features. Multiple imputation models were used.

Discussion

This is the first study to examine the association of marital status on EO-CRC survival and late stage at diagnosis that also compared 5-year CRC survival rates between five racial/ethnic groups using the most recent cancer registries from the SEER program. We found that being unmarried conferred a survival disadvantage regardless of race/ethnicity. In particular, NHB patients who were unmarried experienced the lowest 5-year survival rate (60.3%) than other racial/ethnic groups in our analysis. Consistent with prior literature, NHB patients generally had poorer survival than NHW patients^4,32. Marital status may further exacerbate these disparities. Unmarried patients often have reduced social support, which has been associated with delays in diagnosis, lower adherence to treatment, and poorer overall cancer outcomes^17,19. When combined with structural barriers that disproportionately affect NHB patients—such as limited access to high-quality cancer care, lower rates of CRC screening and symptom awareness³³, and lower likelihood of receiving recommended treatments including adjuvant chemotherapy and radiation therapy^34–38—the absence of marital support may compound existing inequities. These intersecting factors may contribute to the markedly lower survival observed among unmarried NHB patients in our study. Further, health insurance coverage and the presence of chronic disease conditions (e.g., obesity or diabetes)^36–38, may also contribute to the disparities in survival outcomes, particularly for those with more advanced diagnosis. Our imputation results also confirmed this phenomenon. We found that NHB patients were 36% more likely to die from CRC and 7% more likely to have a late stage at diagnosis than NHW patients. When comparing our results with literature that focused on the impact of marital status, our findings align with prior studies²⁷. We found that unmarried NHB patients were 28% more likely to die from CRC and 23% more likely to have a late stage at diagnosis than those who were married. Similarly, Aloysius and colleague reported that being unmarried had a disadvantage in EO-CRC survival in NHB groups²⁷.

In addition, being Asian/PI and Hispanic were also associated with greater risk of death from CRC than NHW patients by 6%-12%. In line with prior literature, Zaki and colleague found that EO-CRC patients were associated with increased all-cause mortality in Asian and Hispanic groups by 6%-16%. Unmarried status also impacted CRC mortality for Asian/PI and Hispanic patients by 10%-17% and was associated with more advanced diagnosis for Hispanic patients by 18% in our analysis. Because there is no study examining the association of marital status on EO-CRC outcomes in Hispanic and Asian/PI groups, it is impossible to directly compare our results with prior literature. Yet, the association of marital status on cancer outcomes has been reported in several studies. Krajic and colleagues found that being unmarried is strongly associated with worse cancer-specific survival, including CRC^39,40. Another study specific to CRC from Saudi Arabia found that unmarried patients were 52% more likely to have an advanced stage and 30% more likely to die from CRC. However, Aloysius and colleagues examined the impact of race and marital status on EO-CRC survival and found that being unmarried had a survival disadvantage in Hispanic, NHB, and NHW patients but not for other racial/ethnic groups, which is partially consistent with our findings. Different results may be because they limited their study sample to those with adenocarcinoma histological subtype and non-US patients with cultural differences in marriage. They also did not further examine the association between marital status, survival, and stage at diagnosis when adjusting for important confounders (e.g., sociodemographic characteristics, tumor features). More importantly, we did not observe significant differences among some racial minority groups, such as AI/AN patients. This likely reflects the limited number of cases reported for these populations within SEER rather than an absence of disparities. Although the racial distribution in our cohort is consistent with prior studies using cancer registry data^41,42, future research that incorporates larger and more diverse datasets will be essential for more accurately characterizing cancer outcomes among underrepresented racial and ethnic groups.

Another notable finding is that Hispanic patients were the youngest at diagnosis in our cohort, followed by AI/AN patients. These patterns are consistent with prior reviews and may reflect broader population age structures, as both Hispanic and AI/AN communities tend to present with younger ages at diagnosis for CRC compared with NHW populations^43,44. These observations highlight the need for further research to clarify the factors contributing to age-related differences in EOCRC presentation across racial and ethnic subgroups. Although our results regarding the association of marital status among NHW patients also consistent with prior literature²⁷, we found that unmarried racial minorities consistently, including NHB and Asian/PI patients, showed the lower survival rates, increased risk of death, and had a greater likelihood of having late stage at diagnosis for CRC. The disparities of CRC survival and stage at diagnosis across these groups may be explained by lack of support, which is closely linked to marital status. Evidence has shown that lack of social support and financial support (including insurance coverage) are generally associated with decreased receipt of any treatment and increased CRC mortality⁴⁵. Young persons from minoritized backgrounds are also more likely to live in disadvantaged neighborhoods with lack of support networks⁴⁶, which may further exacerbate the challenges faced by unmarried individuals. Such evidence highlights the importance of social support on CRC survival and late stage at diagnosis may exist within these young racial minorities.

The main strength of this study is that we considered the association of marital status on EO-CRC survival and late stage at diagnosis, which may reflect how social support influences EO-CRC outcomes. We also examined EO-CRC survival across diverse racial/ethnic groups, which may guide the prevention strategies aimed at prioritizing high-risk groups. Cultural norms—such as typical age at marriage, partnering expectations, and social support—likely contribute to subgroup differences in marital status and may affect both being married at diagnosis and available support during illness. SEER does not capture cultural variables or age at marriage; therefore, our interpretation is hypothesis-generating. Future studies that incorporate sociocultural measures could clarify these pathways. This information is vital to inform culturally sensitive interventions aimed at promoting equitable access to cancer diagnosis/treatment resources for young individuals, across racial/ethnic identities, who may be impacted by lack of social support. Multi-level interventions aimed at enhancing EO-CRC risk awareness and a healthy lifestyle through their social support networks may also be helpful for younger populations. Finally, although we used MI approaches to reduce potential bias due to larger proportion of unknown marital status, future research should also make efforts to improve data completion on reducing unknown marital status for cancer disparity research given that the marital status plays a critical role for social support.

Limitations

First, because it is a retrospective analysis, it relies on existing records and is inherently constrained by the accuracy and completeness of those data. Marital status was captured only as a single snapshot at the time of cancer diagnosis, providing no insight into recent changes. For example, whether an individual married just a month before diagnosis versus having been married for many years, factors that may reflect very different levels of social support. Additionally, marital status serves only as an indirect proxy for psychosocial impact; it represents a readily identifiable characteristic in clinical settings but should not be interpreted as a comprehensive measure of risk or support. Further, another limitation is that more than 20%–25% of patients had missing marital status information, which may introduce bias and weaken the robustness of the findings. We addressed this issue using multiple imputations, and the consistency of results suggests this approach likely reduced bias. More importantly, we were unable to consider other important confounders, such as lifestyle factors (e.g., physical activity, dietary pattern, smoking behavior, alcohol use), comorbidities (e.g., cerebrovascular disease), and systematic barriers (e.g., transportation barriers, lack of treatment or screening facilities), which may have potential impact on cancer outcomes. Finally, we used specialized SEER data (2006–2020) obtained through a formal data-request process that provides census-tract–level information to examine the association of interest. Future research using more recent SEER releases, once sufficient follow‑up time becomes available, will be important to validate our findings and to determine whether the disparities identified in this study persist in more contemporary patient cohorts.

Conclusion

The lowest 5-year CRC survival rate was observed among unmarried EO-CRC patients regardless of racial/ethnic groups, with a significantly lowest survival rate for NHB patients. Culturally sensitive interventions that promote equitable access to health-related care and strengthen social support networks are needed across all racial groups. In particularly, unmarried NHW, NHB and Asian/PI patients were more likely to die from CRC and had more advanced diagnosis. Use of multi-level interventions aimed at improving EO-CRC survival and reducing advanced diagnosis through enhancing their social support networks and CRC risk awareness may also be helpful for these young racial/ethnic groups.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Abbreviations

EO-CRC

Early-onset colorectal cancer

NHW

Non-hispanic white

NHB

Non-hispanic black

AI/AN

American indian/alaskan native

PI

Pacific islanders

SEER

Surveillance, epidemiology, and end results

HR

Hazard ratio

OR

Odd ratio

CI

Confidence interval

Author contributions

Dr. Tsai 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.Concept and design: All authors.Acquisition, analysis, or interpretation of data: Tsai, Coughlin, Gishe.Drafting of the manuscript: Tsai, Coughlin.Critical revision of the manuscript for important intellectual content: All authors.Statistical analysis: Tsai.Obtained funding: Tsai.Administrative, technical, or material support: Coughlin, Gishe, Islam, Eberth.Supervision: Eberth.

Funding

Dr. Meng-Han Tsai was supported by at least in part through the Georgia Cancer Center Paceline funding mechanism at Augusta University (MCGFD01071) and National Cancer Institute (NCI, R21CA301113). The sponsor did not contribute to study design, data collection/analysis/interpretation, manuscript drafting, or manuscript submission. No other conflicts of interest or financial disclosures were reported by the authors of this paper.

Data availability

The datasets generated during the current study are available in the Surveillance, Epidemiology, and End Results Program ( [https://seer.cancer.gov/](https:/seer.cancer.gov) ) repository.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

Data extracted for this study was publicly available and de-identified, and thus considered exempt from IRB review at Augusta University.