ABSTRACT
Objective
To assess how diagnostic pathways affect clinical outcomes and healthcare costs among older adults diagnosed with colorectal cancer (CRC).
Summary of Background Data
Many CRC cases in the United State are diagnosed during an emergency presentation, which have been linked to worse outcomes. Understanding how diagnostic routes may impact a cancer diagnosis may be critical for improving care.
Methods
We analyzed SEER‐Medicare data (2005–2019) for patients aged ≥ 65 years with CRC. Diagnostic routes were classified as screening, inpatient/outpatient (IP/OP), or emergency presentation (ER) based on encounters within 30‐day to 6‐months before diagnosis. Multivariable regression evaluated associations between route and outcomes.
Results
Among 79,251 CRC patients (colon: 72.4%, rectum: 27.6%), diagnoses occurred via ER (13.7%), IP/OP (60.6%), and screening (25.7%). Compared with screened patients, ER‐diagnosed patients were older (median age: 79 vs. 75), had higher comorbidity (22.5% vs. 8.1%), and lower income (29.1% vs. 23.9%). ER diagnosis was associated with urgent surgery (59.7% vs. 20.1%), complications (29.8% vs. 14.2%), extended stay (37.2% vs. 14%), inpatient mortality (7.4% vs. 2.1%), lower discharge to home (32.2% vs. 62.4%), and higher 90‐day mortality (16.8% vs. 5%). Multivariable analysis confirmed ER diagnosis as the strongest predictor of stage IV disease (OR 1.75, 95% CI: 1.66–1.84), urgent surgery (OR 5.14, 95%CI: 4.83–5.48), and 90‐day mortality (OR 1.80, 95%CI: 1.62–1.99; all p < 0.001). Costs were higher for ER versus screening: index surgery ($13,802 vs. $11,458) and postoperative care ($9,470 vs. $3,655).
Conclusions
Emergency CRC diagnosis was associated with worse outcomes and higher costs, emphasizing the need for targeted screening and earlier detection strategies.
Keywords: CRC, emergency diagnosis, healthcare costs, older adults, screening
1. Introduction
Despite transformative advances in preventive medicine and surgical oncology, colorectal cancer (CRC) remains a significant public health and economic burden in the United States [1, 2, 3]. CRC is the fourth most commonly diagnosed malignancy and the second leading cause of cancer‐related mortality among both men and women [1, 2, 3, 4]. By 2025, an estimated 154,270 new CRC cases and 52,900 related deaths are expected [1]. In 2020, CRC accounted for $24.3 billion, or 11.6% of all cancer‐related healthcare costs [5]. Nearly 90% of diagnoses occur among individuals aged 50 + , with incidence rates rising sharply after 65, making older adults especially vulnerable [6]. These trends underscore the urgent need for sustained efforts in early detection, prevention, and system‐level improvements.
National guidelines now recommend initiating CRC screening at age 45, with continuation based on life expectancy, comorbidities, and prior screening history [7, 8]. Medicare also covers approved screening modalities for eligible adults aged 65 and older. Screening uptake among adults aged 45–49 rose 62% from 2019 to 2023, including a 43% increase in colonoscopy and a fivefold rise in stool testing [7, 9, 10]. Despite these gains, approximately one‐third of CRC cases still present emergently, often at advanced stages, raising questions about the reach and screening strategies efficacy [11, 12, 13]. Emergency diagnosis is consistently associated with worse outcomes and elevated costs, regardless of cancer stage [5, 11, 12, 13, 14]. These missed opportunities are especially consequential for Medicare beneficiaries, prompting inquiry into whether increased screening among younger adults has improved outcomes for older populations.
CRC may be diagnosed through multiple clinical pathways, each reflecting varying urgency, resource use, and costs [14, 15, 16]. Research reveals these pathways are not randomly distributed but are deeply shaped by social determinants of health, including race, ethnicity, geography, education, culture, and socioeconomic status [17, 18, 19]. Rural populations and racial and ethnic minorities are disproportionately affected by delayed diagnoses, suggesting late‐stage CRC is not merely a clinical failure but a preventable public health inequity [20, 21].
Systemic barriers‐fragmented care, geographic isolation, limited access to primary care, and inconsistent policy implementation‐contribute to diagnostic delays and worse outcomes [20, 21, 22]. Understanding how patients arrive at diagnosis is essential for identifying missed opportunities and guiding targeted, evidence‐based screening strategies. Despite its relevance, literature on diagnostic routes and their impact on older adults remains insufficiently characterized. This study addresses that gap by leveraging SEER‐Medicare data to examine how diagnostic pathways influence disease stage at presentation, treatment intensity, survival outcomes, and healthcare expenditures. Unlike prior administrative studies, the current analysis applied a validated route‐to‐diagnosis framework across a contemporary 15‐year period, integrated detailed cost trajectories across multiple phases of care, and focused specifically on older Medicare beneficiaries—an understudied population in whom guideline flexibility complicates interpretation of screening “missed opportunities.”
2. Materials and Methods
2.1. Study Design, Data Sources, and Cohort Selection
Patients aged 66–90 years diagnosed with first primary colon and rectum/rectosigmoid cancer (2005–2019) were identified from SEER‐Medicare database using ICD‐O‐3 [23, 24]. SEER aggregates data from 18 registries across 15 states, covering ~ 47.9% of the U.S. population and provides limited patient level demographics, tumor characteristics, treatment modalities and survival outcomes [25]. Medicare covers 97% of adults ≥ 65 and supplies claims data on utilization, screening and diagnostic services, procedures, and providers [26]. SEER‐Medicare linkage achieves a 93% match rate. Eligible participants were continuously enrolled in Medicare Parts A and B under fee‐for‐service for the 12 months before or after diagnosis (or until death). Those enrolled in Medicare Advantage plans, diagnosed by death certificate/autopsy, or, with incomplete staging or benign/in situ cancers were excluded. Cancer‐directed resections were identified using validated ICD‐10‐PCS and CPT codes, diverting procedures without tumor resection were excluded (Supplementary Table 1). For multiple operations, the earliest procedure was considered the index resection. The Institutional Review Board at Ohio State University approved this study and waived the requirement for informed consent, because the data were limited.
2.2. Covariates and Variable Construction
Patient‐level covariates included demographic characteristics (age at diagnosis, sex, race/ethnicity [categorized as White, Black, Hispanic, and other], marital status, region, urban/rural status, and Social Vulnerability Index [SVI]), Charlson Comorbidity Index (CCI), cancer site, stage, route to diagnosis (RTD), length of stay, survival, Medicare status, and cost data. Comorbidity burden was assessed using the CCI and categorized as low (≤ 2) or high‐risk (> 2). Socioeconomic vulnerability was evaluated using SVI, based on income, education, housing, and transportation grouped into high, moderate or low using county‐level medians [27, 28]. Income data were available only at area level.
2.3. Exposure Variable: Route to Diagnosis
The primary exposure variable was the route to diagnosis (RTD), defined using a validated algorithm based on healthcare encounters before diagnosis and categorized as Routine Screening, Inpatient/Outpatient (IP/OP) Visits, or Emergency route (ER) [14]. RTD classification followed a hierarchical, stepwise algorithm. First, we assessed whether a patient underwent a CRC screening test within 6 months prior to diagnosis. Encounters within 30 days prior to diagnosis were prioritized. If present, a diagnosis was categorized as screening‐related, prioritizing events occurring within 30 days of diagnosis. For patients without screening, we evaluated healthcare encounters within the preceding 6 months. We started the RTD leading to colorectal cancer detection by examining two primary pathways, based on whether the patient received CRC screening within 6 months prior to diagnosis or not. If screening was documented, the pathway was further divided into diagnoses occurring within 30 days or within 6 months of the screening event. For patients without documented screening in the 6 months prior to the diagnosis, we assessed whether they had any emergency department, inpatient, or outpatient visits within 30 days before diagnosis. Based on the type and timing of these encounters, patients were categorized into the “30‐day ER route,” “30‐day IP/OP route,” or, if no visits occurred in that timeframe, into the “6‐month ER route” or “6‐month IP/OP route.” (Figure 1). Emergency visits and CRC screening services were identified using the appropriate revenue codes and CPT/HCPCS/ICD codes from hospital discharge records, outpatient, MEDPAR, and carrier files [23, 24]. This approach emphasized the most proximate clinical level interaction leading to diagnosis while minimizing misclassification across competing routes.
Figure 1.
Mapping the pathways to CRC detection and healthcare utilization patterns.
2.4. Outcomes of Interest
Primary outcomes assessed RTD's impact on both clinical and economic trajectories. Key endpoints included textbook outcomes (TO)‐ a composite, literature‐defined measure of optimal short‐term surgical outcomes defined by absence of prolonged hospital stay (> 75th percentile), complications, 90‐day readmission, and 90‐day mortality [29]. Additional outcome included delayed surgery, surgery of surgery, discharge disposition, and costs for index surgery, and cumulative expenditures at 90 days and 1 year.
2.5. Statistical Analysis
Descriptive statistics summarized the cohort. Categorical variables were reported as frequencies; continuous variables as means (SD) or medians (IQR). Associations between RTD and categorical variables used Chi‐squared tests, while the Kruskal–Wallis H test used for continuous variables. Multivariable multinomial logistic regression assessed RTD associations with outcomes, adjusting for patient and hospital factors. Bonferroni correction addressed multiple comparisons. Sensitivity analyses stratified outcomes by cancer sites. All tests were two‐sided (p < 0.05). Adjusted odds ratios (aORs) and 95% confidence intervals (CIs) were reported. Analyses used SAS 9.4 and R 4.2.1.
3. Results
3.1. Baseline Patient Characteristics
A total of 79,251 Medicare beneficiaries aged 66–90 years underwent CRC surgery for a malignant indication. Of these, 77.6% had colon cancer (n = 61,222) and 22.4% had rectal cancer (n = 18,029). Median age was 77 years (IQR: 72–83), and most were female (n = 43,291, 54.4%). Most individuals self‐identified as White (n = 63,595, 80%), followed by Black (n = 6,682, 8.4%), Hispanic (n = 4,920, 6.6%), and other race/ethnicities (n = 4,324, 5.4%). Geographically, most individuals resided in metropolitan areas (n = 65,452, 82.3%), with highest representation from the West (n = 36,083, 45.5%) or South (n = 31,025, 39%). Diagnostic pathways varied: IP/OP visits (n = 48,305, 60.1%), routine screening (n = 20,366, 25.7%), and emergency presentation (n = 10,850, 13.7%) (Table 1). About half of patients were treated for advanced‐stage disease (stage III or IV) (n = 38,234, 48.1%). Most individuals underwent elective surgery (n = 37,845, 62.1%). Overall, 50.3% (n = 30,662) of the cohort had surgery at non‐teaching hospitals, while 49.7% (n = 30,332) were treated at teaching hospitals.
Table 1.
Baseline characteristics of patients.
| Characteristics | Total (n = 79,251) | Emergency route (ER) (n = 10,850; 13.7%) | In‐patient/outpatient visits (n = 48,305.60.6%) | Routine screening (n = 20,366; 25.7%) | p value |
|---|---|---|---|---|---|
| Age, years | 77 (72, 83) | 79 (73, 84) | 78 (72, 83) | 75 (71, 81) | < 0.001 |
| Sex | < 0.001 | ||||
| Male | 36,230 (45.6%) | 4,743 (43.7%) | 21,880 (45.3%) | 9,607 (47.2%) | |
| Female | 43,291 (54.4%) | 6,107 (56.3%) | 26,425 (54.7%) | 10,759 (52.8%) | |
| CCI | < 0.001 | ||||
| ≤ 2 | 71,275 (89.6%) | 8,413 (77.5%) | 44,151 (91.4%) | 18,711 (91.9%) | |
| > 2 | 8,246 (10.4%) | 2,437 (22.5%) | 4,154 (8.6%) | 1,655 (8.1%) | |
| Race | < 0.001 | ||||
| White | 63,595 (80%) | 8,346 (76.9%) | 38,747 (80.2%) | 16,502 (81%) | |
| Black | 6,682 (8.4%) | 1,243 (11.5%) | 3,869 (8%) | 1,570 (7.7%) | |
| Hispanic | 4,920 (6.2%) | 738 (6.8%) | 3,003 (6.2%) | 1,179 (5.8%) | |
| Other | 4,324 (5.4%) | 523 (4.8%) | 2,686 (5.6%) | 1,115 (5.5%) | |
| Racial minority | < 0.001 | ||||
| Non‐minority | 63,595 (80%) | 8,346 (76.9%) | 38,747 (80.2%) | 16,502 (81%) | |
| Minority | 15,926 (20%) | 2,504 (23.1%) | 9,558 (19.8%) | 3,864 (19%) | |
| Region | < 0.001 | ||||
| Midwest | 6,697 (8.4%) | 1,029 (9.5%) | 3,994 (8.3%) | 1,674 (8.2%) | |
| Northeast | 5,716 (7.2%) | 865 (8%) | 3,390 (7%) | 1,461 (7.2%) | |
| South | 31,025 (39%) | 4,242 (39.1%) | 19,154 (39.7%) | 7,629 (37.5%) | |
| West | 36,083 (45.4%) | 4,714 (43.4%) | 21,767 (45.1%) | 9,602 (47.1%) | |
| Metro | < 0.001 | ||||
| Rural | 14,065 (17.7%) | 2,139 (19.7%) | 8,131 (16.8%) | 3,795 (18.6%) | |
| Urban | 65,452 (82.3%) | 8,710 (80.3%) | 40,172 (83.2%) | 16,570 (81.4%) | |
| Married | < 0.001 | ||||
| Married | 37,398 (47%) | 5,955 (54.9%) | 23,049 (47.7%) | 8,394 (41.2%) | |
| Not married | 38,400 (48.3%) | 4,366 (40.2%) | 23,046 (47.7%) | 10,988 (54%) | |
| Income quintile | < 0.001 | ||||
| Quintile I | 18,324 (25%) | 2,951 (29.1%) | 10,873 (24.5%) | 4,500 (23.9%) | |
| Quintile II | 18,300 (25%) | 2,616 (25.8%) | 11,083 (25%) | 4,601 (24.4%) | |
| Quintile III | 18,326 (25%) | 2,413 (23.8%) | 11,146 (25.2%) | 4,767 (25.3%) | |
| Quintile IV | 18,298 (25%) | 2,156 (21.3%) | 11,191 (25.3%) | 4,951 (26.3%) | |
| SVI category | < 0.001 | ||||
| Low | 26,479 (33.3%) | 3,491 (32.2%) | 16,265 (33.7%) | 6,723 (33%) | |
| Moderate | 26,539 (33.4%) | 3,551 (32.7%) | 15,980 (33.1%) | 7,008 (34.4%) | |
| High | 26,495 (33.3%) | 3,807 (35.1%) | 16,054 (33.2%) | 6,634 (32.6%) | |
| Cancer site | < 0.001 | ||||
| Colon | 61,492 (77.3%) | 8,727 (80.4%) | 37,143 (76.9%) | 15,622 (76.7%) | |
| Rectum | 18,029 (22.7%) | 2,123 (19.6%) | 11,162 (23.1%) | 4,744 (23.3%) | |
| Cancer stage | < 0.001 | ||||
| I | 18,626 (23.4%) | 2,050 (18.9%) | 10,319 (21.4%) | 6,257 (30.7%) | |
| II | 22,661 (28.5%) | 3,134 (28.9%) | 13,815 (28.6%) | 5,712 (28%) | |
| III | 21,293 (26.8%) | 2,812 (25.9%) | 12,976 (26.9%) | 5,505 (27%) | |
| IV | 16,941 (21.3%) | 2,854 (26.3%) | 11,195 (23.2%) | 2,892 (14.2%) | |
| Teaching status | 0.004 | ||||
| Non‐teaching | 30,662 (50.3%) | 3,850 (49.1%) | 18,585 (50.8%) | 8,227 (49.7%) | |
| Teaching | 30,332 (49.7%) | 3,993 (50.9%) | 18,000 (49.2%) | 8,339 (50.3%) | |
| Surgery type | < 0.001 | ||||
| Elective | 37,845 (62.1%) | 3,155 (40.3%) | 21,466 (58.7%) | 13,224 (79.9%) | |
| Non‐elective | 23,078 (37.9%) | 4,677 (59.7%) | 15,076 (41.3%) | 3,325 (20.1%) |
Note: Significance is based on p < 0.05.
Abbreviations: CCI, Charlson Comorbidity Index; ER, emergency route; IP, inpatient; OP, outpatient; SVI, Social Vulnerability Index.
Patients diagnosed via emergency presentation were older (median patient age 79 years vs. 75 years; p < 0.001), had a high (> 2) CCI score (22.5% vs. 8.1%; p < 0.001), and were more likely to present with Stage IV disease (26.3% vs. 14.2%; p < 0.001) versus individuals diagnosed through screening. Notably, screened patients were more often diagnosed at Stage I compared with patients diagnosed through IP/OP and ER routes (30.7% vs. 21.4% vs. 18.9%; p < 0.001). Patients diagnosed through ER route were more likely to be in the lowest income quintile compared with patients diagnosed with IP/OP and screening routes (29.1% vs. 24.5% vs. 23.9%; p < 0.001). Similarly, patients screened through ER route were more likely to be from minority races versus patients diagnosed through IP/OP and screening routes, respectively (23.1% vs. 19.8% vs. 19%; p < 0.001). Colon cancer was slightly more prevalent among ER cases than IP/OP routes and screening routes (80.4% vs. 76.9%, 76.7%; p < 0.001) (Table 1).
3.2. Temporal Trends in CRC Diagnosis Route
Over the 15‐year study period, distribution of CRC diagnosis routes shifted notably (Figure 2). Between 2005 and 2019, the proportion of colorectal cancer diagnoses via routine screening increased from 25% to 35%, while diagnoses through IP/OP evaluation routes declined from 64% to 50%. Emergency presentations remained relatively stable, ranging from 11% to 15% across all years. By the end of the study period, IP/OP was the most commonly used diagnosis route, followed by screening.
Figure 2.
Route to CRC diagnosis: ER, IP/OP and Routine Screening (2005‐2019).
3.3. Factors Associated With ER and IP/OP Diagnosis Routes
Multivariable modeling of factors influencing emergency and IP/OP diagnostic routes in colorectal cancer revealed associations with sociodemographic and clinical characteristics. Racial minority status was a strong predictor of diagnosis through ER (aOR 1.34; 95%CI: 1.25–1.42; p < 0.001) and IP/OP (aOR 1.10; 95%CI: 1.05–1.15; p < 0.001) routes. Lower income levels were similarly associated with higher odds of diagnosis through ER and IP/OP routes. Patients in the lowest income quintile had 46% higher odds of emergency diagnosis (aOR 1.46; 95%CI: 1.35–1.58; p < 0.001) and 14% higher odds of inpatient/outpatient diagnosis (aOR 1.14; 95%CI: 1.08–1.20; p < 0.001) compared with patients in the highest income quartile. Geographic disparities were also evident, with individuals residing in the South exhibiting increased odds of both emergency presentation (aOR 1.16; 95%CI: 1.09–1.23) and IP/OP diagnosis (aOR 1.12; 95%CI: 1.08–1.17) than the West region (both p < 0.001). Clinical factors contributed substantially, with a higher comorbidity index (CCI > 2) strongly associated with emergency diagnosis (aOR 2.98; 95%CI: 2.78–3.20; p < 0.001). Additionally, older age and earlier year of diagnosis were associated with ER and IP/OP routes. These findings underscore the multifactorial nature of diagnostic disparities and highlight structural and clinical determinants on colorectal cancer detection pathways (Table 2).
Table 2.
Risk factors associated with ER and IP/OP diagnosis routes.
| Variable | RTD | Adjusted odds ratio | 95% CI | p value |
|---|---|---|---|---|
| Age at diagnosis | ER | 1.05 | 1.04–1.06 | < 0.001 |
| IP/OP | 1.04 | 1.03–1.05 | < 0.001 | |
| CCI > 2 | ER | 2.98 | 2.78–3.20 | < 0.001 |
| IP/OP | 1.03 | 0.94–1.07 | 0.933 | |
| Racial minority | ER | 1.34 | 1.25–1.42 | < 0.001 |
| IP/OP | 1.10 | 1.05–1.15 | < 0.001 | |
| Income | ||||
| Quintile I | ER | 1.46 | 1.35–1.58 | < 0.001 |
| IP/OP | 1.14 | 1.08–1.20 | < 0.001 | |
| Quintile II | ER | 1.31 | 1.21–1.41 | < 0.001 |
| IP/OP | 1.12 | 1.07–1.18 | < 0.001 | |
| Quintile III | ER | 1.16 | 1.08–1.25 | < 0.001 |
| IP/OP | 1.06 | 1.00–1.11 | 0.032 | |
| Region | ||||
| Midwest | ER | 1.13 | 1.03–1.24 | 0.007 |
| IP/OP | 1.01 | 0.95–1.08 | 0.725 | |
| Northeast | ER | 1.08 | 0.97–1.19 | 0.151 |
| IP/OP | 1.01 | 0.94–1.08 | 0.793 | |
| South | ER | 1.16 | 1.09–1.23 | < 0.001 |
| IP/OP | 1.12 | 1.08–1.17 | < 0.001 | |
| Year of diagnosis | ER | 0.98 | 0.98–0.99 | < 0.001 |
| IP/OP | 0.95 | 0.94–0.96 | < 0.001 |
Note: Significance is based on p < 0.05.
Abbreviations: CCI, Charlson Comorbidity Index; CI, confidence interval; ER, emergency route; IP, inpatient; OP, outpatient.
3.4. Short Term Clinical and Economic Outcomes by CRC Route of Diagnosis
Compared with patients diagnosed through routine screening, individuals diagnosed via the ER experienced markedly different outcomes. Emergency route patients were more likely to undergo urgent surgery (59.7% vs. 20.1%, p < 0.001), but were less likely to receive surgery overall (74.5% vs. 85%, p < 0.001). These individuals had extended hospital stays compared with patients diagnosed through routine screening (37.2% vs. 14%, p < 0.001). They also had a higher postoperative complication (29.8% vs. 14.2%, p < 0.001). In‐hospital mortality was higher among emergency route patients versus patients with routine screening route (7.4% vs. 2.1%, p < 0.001). Furthermore, only 32.2% of emergency route patients were discharged to home, in contrast to 62.4% of patients in the screening group (p < 0.001). A greater proportion of ER route patients required discharge to a skilled nursing facility compared with patients diagnosed through screening (32.7% vs. 13.8%, p < 0.001), reflecting the increased acuity and complexity of their clinical course (Table 3).
Table 3.
Clinical outcomes and expenditure burden by diagnostic pathway.
| Variable name | Total (n = 79,251) | ER (n = 10,850,13.7%) | IP/OP (n = 48,305,60.6%) | Screening (n = 20,366,25.7%) | p value |
|---|---|---|---|---|---|
| Chemotherapy | 44,903 (56.5%) | 5,861 (54%) | 26,914 (55.7%) | 12,128 (59.6%) | < 0.001 |
| Radiotherapy | 10,624 (13.4%) | 1,173 (10.8%) | 6,738 (13.9%) | 2,713 (13.3%) | < 0.001 |
| Surgery | 63,839 (80.3%) | 8,088 (74.5%) | 38,434 (79.6%) | 17,317 (85%) | < 0.001 |
| Urgent surgery | 23,078 (37.9%) | 4,677 (59.7%) | 15,076 (41.3%) | 3,325 (20.1%) | < 0.001 |
| Mean time to surgery day | 11 (0, 30) | 0 (0, 20) | 9 (0, 29) | 19 (3, 36) | < 0.001 |
| Textbook outcome | 27,794 (45.6%) | 2,158 (27.5%) | 16,016 (43.8%) | 9,620 (58.1%) | < 0.001 |
| Any complication | 12,129 (19.9%) | 2,339 (29.8%) | 7,434 (20.3%) | 2,356 (14.2%) | < 0.001 |
| LOS > 75th percentile | 13,999 (23%) | 2,920 (37.2%) | 8,766 (24%) | 2,313 (14%) | < 0.001 |
| In‐hospital death | 2,348 (3.8%) | 579 (7.4%) | 1,416 (3.9%) | 353 (2.1%) | < 0.001 |
| 90‐day Readmission | 23,950 (39.3%) | 4,098 (52.3%) | 14,777 (40.4%) | 5,075 (30.6%) | < 0.001 |
| 90‐day Mortality | 5,603 (9.2%) | 1,315 (16.8%) | 3,460 (9.5%) | 828 (5%) | < 0.001 |
| Discharge disposition | < 0.001 | ||||
| Home | 30,190 (49.5%) | 2,528 (32.2%) | 17,323 (47.4%) | 10,339 (62.4%) | |
| SNF | 13,322 (21.8%) | 2,567 (32.7%) | 8,464 (23.1%) | 2,291 (13.8%) | |
| Hospice | 686 (1.1%) | 158 (2%) | 435 (1.2%) | 93 (0.6%) | |
| HHA | 12,340 (20.2%) | 1,567 (20%) | 7,686 (21%) | 3,087 (18.6%) | |
| 1‐year pre‐surgery expenditures ($) | 1,378 (277–5,724) | 4,476 (1,035–12,670) | 1,063 (140–5,031) | 1,370 (422–4,423) | < 0.001 |
| Index surgery ($) | 12,125 (92,245–15392) | 13,802 (11,159–23,424) | 12,130 (9,406–15,624) | 11,458 (8,409–13,685) | < 0.001 |
| 90‐day post‐surgery expenditures ($) | 1,292 (359–8,579) | 3,638 (81–12,788) | 1,363 (36–8,761) | 719 (21–6,115) | < 0.001 |
| 1‐year post‐surgery expenditures ($) | 5,431 (612–18,740) | 9,470 (1,131–26,432) | 5,605 (616–18,769) | 3,655 (532–15,207) | < 0.001 |
Note: Significance is based on p < 0.05.
Abbreviations: ER, emergency route, HHA, home health agency; IP, inpatient; LOS, Length of hospital stay; OP, outpatient; SNF, skilled nursing facility.
After surgery, patients diagnosed through the ER route, compared with individuals diagnosed via routine screening, had higher 90‐day readmission rates (emergency route: 52.3% vs. routine screening: 30.6%, p < 0.001), and were less likely to achieve textbook outcomes (emergency route: 27.5% vs. routine screening: 58.1%, p < 0.001). In addition, patients diagnosed through the emergency route incurred substantially higher inflation‐adjusted expenditures (all costs standardized to 2019 U.S. dollars using the Medicare Economic Index) across all phases of care compared with both the screening and IP/OP groups. These expenditures included higher pre‐surgical costs ($4,476 vs. $1,370), index surgery costs ($13,802 vs. $11,458), 90‐day post‐surgery costs ($3,638 vs. $719), and 1‐year post‐surgery expenses ($9,470 vs. $3,655) relative to routine screening diagnoses (all p < 0.001). Furthermore, 90‐day mortality was threefold higher in the emergency group than the screening route (emergency route: 16.8% vs. screening route: 5%, p < 0.001) (Table 3). These results were consistent in stratified analyses by cancer site (Supplementary Table 2).
3.5. Multivariable Logistic Regression Models Assessing Clinical and Financial Outcomes
Multivariable analysis revealed persistent health and cost disparities by patient diagnosis route, with emergency route associated with the worst outcomes, followed by IP/OP routes, compared with routine screening. The odds of delayed diagnosis were notably higher among ER route groups (aOR: 1.75, 95%CI: 1.66–1.84, p < 0.001), and IP/OP groups (aOR: 1.53, 95%CI: 1.48–1.59, p < 0.001) versus screening. Furthermore, ER (aOR: 0.48, 95%CI: 0.45–0.52, p < 0.001) and IP/OP patients (aOR: 0.81, 95%CI: 0.77–0.85, p < 0.0001) had ower odds of receiving surgical treatment with ER patients being the least likely to undergo surgery. Conversely, the likelihood of undergoing urgent surgery was higher for ER (aOR: 5.14, 95%CI: 4.83–5.48, p < 0.001), followed by IP/OP patients (aOR: 2.76, 95%CI: 2.63–2.90, p < 0.001) compared with patients diagnosed through screening. Complication rates were elevated in both ER (aOR: 1.55, 95%CI: 1.44–1.67, p < 0.001) and IP/OP groups (aOR: 1.22, 95%CI: 1.16–1.29, p < 0.001); and extended hospital stays were higher (ER: aOR: 1.81, 95%CI: 1.69–1.95; IP/OP: aOR: 1.34, 95%CI: 1.27–1.42; both p < 0.001) in ER, followed by IP/OP groups compared with patients diagnosed through screening. Patients diagnosed via ER were the least likely to be discharged home (aOR: 0.55, 95%CI: 0.51–0.59 p < 0.001), followed by IP/OP patients (aOR: 0.76, 95%CI: 0.73–0.80, p < 0.001). Hospital mortality was higher among ER (aOR: 1.75, 95%CI: 1.50–2.03, p < 0.001), followed by IP/OP patients (aOR: 1.24, 95%CI: 1.09–1.41, p = 0.001), and 90‐day mortality followed a similar trend (ER: aOR: 1.80, 95%CI: 1.62–2.00; IP/OP: aOR: 1.29, 95%CI: 1.18–1.41; both p < 0.001). Achievement of textbook outcomes was lower among ER (aOR: 0.53, 95%CI: 0.50–0.57, p < 0.001), followed by IP/OP patients (aOR: 0.79, 95%CI: 0.75–0.82, p < 0.001), compared to patients diagnosed through screening. Financial burden was markedly higher across all phases of care for ER patients compared with screened patients, including index expenditure (ER: 14.66% increase, 95%CI: 12.98–16.35, p < 0.001), 90‐day post‐surgery expenditure (32.05% increase, 95%CI: 25.99–38.40, p < 0.001), 1‐year post‐surgery expenditure (30% increase, 95%CI: 24.8–35.41, p < 0.001), and total pre‐surgery expenditure (65.18% increase, 95%CI: 59.08–71.51, p < 0.001). IP/OP patients also faced elevated costs, with index expenditure rising by 6.91% (95%CI: 5.85–7.97, p < 0.001), 90‐day post‐surgery costs by 12.73% (95%CI: 9.15–16.43, p < 0.001), 1‐year post‐surgery expenses by 11.14% (95%CI: 8.13–14.23, p < 0.001), and total pre‐surgery costs by 20.21% (95%CI: 17.11–23.4, p < 0.001). These findings underscore the substantial clinical and economic burden non‐screening colorectal cancer diagnosis routes (Table 4 and Figure 3).
Table 4.
Multivariable analysis of the health and cost impact of emergency and inpatient/outpatient diagnosis routes compared to routine CRC screening.
| Variable | Adjusted odds ratio/percent difference | 95% CI (Lower‐upper) | p value |
|---|---|---|---|
| Delayed diagnosis | |||
| ER | 1.75 | 1.66–1.84 | < 0.001 |
| IP/OP | 1.53 | 1.48–1.59 | < 0.001 |
| Receipt of surgery | |||
| ER | 0.48 | 0.45–0.52 | < 0.001 |
| IP/OP | 0.81 | 0.77–0.85 | < 0.001 |
| Urgent surgery | |||
| ER | 5.14 | 4.83–5.48 | < 0.001 |
| IP/OP | 2.76 | 2.63–2.89 | < 0.001 |
| Textbook outcome | |||
| ER | 0.53 | 0.50–0.57 | < 0.001 |
| IP/OP | 0.79 | 0.75–0.82 | < 0.001 |
| Any complication | |||
| ER | 1.55 | 1.44–1.67 | < 0.001 |
| IP/OP | 1.22 | 1.16–1.29 | < 0.001 |
| Extended length of stay | |||
| ER | 1.81 | 1.69–1.95 | < 0.001 |
| IP/OP | 1.34 | 1.27–1.42 | < 0.001 |
| In‐hospital mortality | |||
| ER | 1.75 | 1.50–2.03 | < 0.001 |
| IP/OP | 1.24 | 1.09–1.41 | 0.001 |
| 90‐day readmission | |||
| ER | 1.55 | 1.45–1.65 | 0.374 |
| IP/OP | 1.20 | 1.15–1.26 | 0.141 |
| 90‐day mortality | |||
| ER | 1.8 | 1.62–2.00 | < 0.001 |
| IP/OP | 1.29 | 1.18–1.41 | < 0.001 |
| Discharge home | |||
| ER | 0.55 | 0.51–0.59 | < 0.001 |
| IP/OP | 0.76 | 0.73–0.80 | < 0.001 |
| Index expenditures | |||
| ER | 14.66 | 12.98–16.35 | < 0.001 |
| IP/OP | 6.91 | 5.85–7.97 | < 0.001 |
| 90‐day expenditures | |||
| ER | 32.05 | 25.99–38.4 | < 0.001 |
| IP/OP | 12.73 | 9.15–16.43 | < 0.001 |
| 1‐year postoperative expenditures | |||
| ER | 30 | 24.8–35.41 | < 0.001 |
| IP/OP | 11.14 | 8.13–14.23 | < 0.001 |
| Total preoperative expenditures | |||
| ER | 65.18 | 59.08–71.51 | < 0.001 |
| IP/OP | 20.21 | 17.11–23.4 | < 0.001 |
Note: Significance is based on p < 0.05.
Abbreviations: CI: confidence interval; ER, emergency route; IP, inpatient; OP, outpatient; TO: textbook outcome.
Figure 3.
Diagnostic route and surgical outcomes.
4. Discussion
Despite decades of emphasis on colorectal cancer screening, emergency diagnoses persist, driving poor outcomes, higher costs, and exposing structural deficiencies in U.S. healthcare [17, 18, 30, 31]. Leveraging SEER‐Medicare data from 2005 to 2019, the current study reinforced exiting literature while offering a nuanced perspective on how diagnostic pathways shape health trajectories and economic burden among older adults. The study findings demonstrated that routine screening diagnoses rose from 25% to 35%, while IP/OP diagnostic encounters declined from 64% to 50%; emergency presentations increased from 11% to 15%. This trend reveals a troubling disconnect: as preventive initiatives expand, vulnerable populations remain systematically excluded from their benefits.
The downstream consequences of delayed diagnosis are profound. Patients diagnosed emergently were markedly more likely to present with advanced‐stage disease, undergo urgent surgery, and represent lower‐income, rural, and racially minoritized groups, as well as individuals with high comorbidity burdens. Notably, these patients had longer hospitalizations, higher in‐hospital and 90‐day mortality, and a 55% grater readmission risk and markedly increased treatment costs. Nearly one‐third required discharge to skilled nursing facilities, and their post‐surgical expenditures were up to five times higher than individuals diagnosed via screening. These disparities are not incidental—they are predictable outcomes of a healthcare system that fails to equitably deliver preventive services.
The findings align with a growing body of evidence implicating limited access to primary care, geographic isolation, fragmented care coordination, and inconsistent policy implementation as a driver of diagnostic delay and poorer clinical outcomes [32, 33, 34, 35]. Lower health literacy and historical mistrust in medical institutions further compound these barriers, particularly among racial and ethnic minorities [36, 37]. For older adults, competing priorities from chronic disease management often overshadow cancer screening [14]. Taken together, emergency CRC diagnosis, therefore, should not be interpreted as a failure of individual compliance, but as a manifestation of systemic neglect, an indictment of missed opportunities for timely detection and coordinated care.
This diagnostic divergence reflects the concept of “structural vulnerability,” wherein health outcomes are shaped less by personal choices and more by institutional arrangements and policy environments [38, 39, 40]. Addressing this issue demands a paradigm shift from reactive, episodic care to proactive, equity‐centered prevention. Coverage expansion alone is insufficient. Strategic investments must confront the social determinants of health through culturally tailored outreach, longitudinal care continuity, and community‐informed program design. Building trust within historically marginalized communities requires workforce diversification, sustained engagement and accountability mechanisms. Moreover, data‐driven tools—such as geospatial mapping and predictive analytics, can optimize resource allocation to high‐risk populations. Financially, the case for reform is compelling: bundled payment models and value‐based care frameworks that reward early detection and penalize avoidable emergency presentations offer a path to both cost containment and improved outcomes. For older adults with complex health needs, ethical surgical practices and timely diagnosis are not just clinical imperatives, they are moral and fiscal necessities.
While this study provides important insights into the relationship between diagnostic pathways and CRC outcomes, several limitations warrant consideration. First, the SEER–Medicare database primarily includes older adults (aged 65–99 years), limiting generalizability to younger populations. Second, the study period (2005–2019) spans evolving screening guidelines, healthcare delivery models, and coding practices, which may introduce selection bias. Third, diagnostic routes were inferred from administrative claims and may not have fully captured clinical nuance, screening intent, or symptom onset. Claims data did not reliably ascertain time since prior screening beyond the 6 months preceding diagnosis nor consistently distinguish screening from diagnostic procedures, precluding assessment of long‐term screening adherence and complicating interpretation of screening cessation among adults aged 76–85 years. More broadly, claims‐based route‐to‐diagnosis classifications reflected healthcare utilization patterns rather than patient‐reported symptoms, clinician intent, or shared decision‐making—factors particularly relevant for older adults for whom screening was guideline‐flexible. Accordingly, non‐screening routes should be interpreted as diagnostic pathways rather than definitive failures of screening adherence. Finally, cost estimates capture direct Medicare expenditures but exclude indirect costs such as caregiver burden and out‐of‐pocket expenses. Despite these limitations, our findings robustly demonstrate that diagnostic route is a key determinant of clinical outcomes and healthcare spending.
5. Conclusions
Emergency CRC diagnosis emerges was the strongest predictor of poor prognosis, disproportionately affecting socioeconomically disadvantaged and medically complex individuals. Reducing these diagnoses require systemic reforms that dismantle structural barriers, expand early detection infrastructure, and reach high‐risk populations through targeted, culturally responsive interventions. Future research should incorporate qualitative methods and broader socioeconomic indicators to illuminate the lived experiences and structural constraints delaying diagnosis. By reframing diagnostic pathways as reflections of systemic inequity, this study calls for a reimagining of cancer care—one that centers equity, anticipates risk, and prioritizes prevention over crisis response.
Funding
The authors have nothing to report.
Conflicts of Interest
The authors declare no conflicts of interest.
Synopsis
This study demonstrates that older adults diagnosed with colorectal cancer following emergency presentations experience worse clinical outcomes, higher mortality rates, and substantially increased healthcare costs compared with individuals diagnosed through routine screening. Emergency diagnoses are strongly associated with advanced disease, greater comorbidity, and socioeconomic disadvantage, reflecting persistent disparities in early detection. These findings underscore the importance of implementing targeted, equity‐focused screening and prevention strategies to reduce avoidable emergency presentations and improve outcomes.
Supporting information
Table S1: Medical Procedure and Screening Codes Reference.
Table S2: Health and Financial Outcomes of Colon and Rectal Cancer Patients by Care Site.
Acknowledgments
The collection of cancer incidence data used in this study was supported by the California Department of Public Health pursuant to California Health and Safety Code Section 103885; Centers for Disease Control and Prevention's (CDC) National Program of Cancer Registries, under cooperative agreement 1NU58DP007156; the National Cancer Institute's Surveillance, Epidemiology and End Results Program under contract HHSN261201800032I awarded to the University of California, San Francisco, contract HHSN261201800015I awarded to the University of Southern California, and contract HHSN261201800009I awarded to the Public Health Institute. The ideas and opinions expressed herein are those of the author(s) and do not necessarily reflect the opinions of the State of California, Department of Public Health, the National Cancer Institute, and the Centers for Disease Control and Prevention or their Contractors and Subcontractors.
Data Availability Statement
The data for this study were obtained from the linked SEER‐Medicare database. There are restrictions to the availability of this data, which is used under license for this study. Data can be accessed with permission from the National Cancer Institute and Center for Medicare and Medicaid Services.
References
- 1. American Cancer Society ., Cancer Facts & Figures 2025 (Atlanta, GA: American Cancer Society, 2025). accessed August 23, 2025, https://www.cancer.org/research/cancer-facts-statistics/all-cancer-facts-figures/2025-cancer-facts-figures.html. [Google Scholar]
- 2. National Cancer Institute . Colorectal Cancer Stat Facts. Surveillance, Epidemiology, and End Results Program, accessed August 24, 2025, https://seer.cancer.gov/statfacts/html/colorect.html.
- 3. Schafer E. J., Sung H., Star J., Bandi P., Smith R. A., and Siegel R. L., “Colorectal Cancer Incidence in US Adults After Recommendations for Earlier Screening,” Journal of the American Medical Association 334 (2025): 824–826, 10.1001/jama.2025.9147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. American Cancer Society . Key statistics for colorectal cancer. Updated April 28, 2025, accessed August 24, 2025, https://www.cancer.org/cancer/types/colon-rectal-cancer/about/key-statistics.html.
- 5. Centers for Disease Control and Prevention . Health and Economic Benefits of Colorectal Cancer Interventions. Updated March 2024, accessed September 16, 2025, https://www.cdc.gov/nccdphp/priorities/colorectal-cancer.html.
- 6. American Cancer Society ., Colorectal Cancer Facts & Figures 2023–2025 (Atlanta, GA: American Cancer Society, 2022). Accessed August 23, 2025, https://www.cancer.org/research/cancer-facts-statistics/colorectal-cancer-facts-figures.html. [Google Scholar]
- 7. Star J., Siegel R. L., Smith R. A., Schafer E. J., Jemal A., and Bandi P., “Trends in Colorectal Cancer Screening in US Adults Aged 45 to 49 Years,” Journal of the American Medical Association 334, no. 9 (2025): 827–829, 10.1001/jama.2025.10618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. U.S. Preventive Services Task Force ., “Screening for Colorectal Cancer: US Preventive Services Task Force Recommendation Statement,” Journal of the American Medical Association 325, no. 19 (2021): 1965–1977, 10.1001/jama.2021.6238. [DOI] [PubMed] [Google Scholar]
- 9. Lee J. K., Jensen C. D., Hendel J., Udaltsova N., Corley D. A., and Levin T. R., “Screening Colonoscopy Yields among Adults Aged 45 to 49 Years After Lowering the Colon Cancer Screening Age,” Journal of the American Medical Association 334, no. 5 (2025): 449–452, 10.1001/jama.2025.7494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Centers for Disease Control and Prevention . Colorectal Cancer Screening Rates. Updated March 2024. Available from: https://www.cdc.gov/cancer/colorectal/statistics/screening_rates.htm.
- 11. Khalaf N., Ali B., Zimolzak A. J., et al., “Prevalence and Outcomes of Emergency Presentations of Colorectal Cancer in Veterans Affairs Health Care System,” Digestive Diseases and Sciences 70 (2025): 177–190, 10.1007/s10620-024-08762-6. [DOI] [PubMed] [Google Scholar]
- 12. Mariotto A. B., Enewold L., Zhao J., Zeruto C. A., and Yabroff K. R., “Medical Care Costs Associated With Cancer Survivorship in the United States,” Cancer Epidemiology, Biomarkers and Prevention 29, no. 7 (2020): 1304–1312, 10.1158/1055-9965.EPI-19-1379. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. American Cancer Society ., Colorectal Cancer Facts & Figures 2020–2022 (Atlanta, GA: American Cancer Society, 2020). accessed September 1, 2025, https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/colorectal-cancer-facts-and-figures-2020-2022.pdf. [Google Scholar]
- 14. Pennisi F., Buzzoni C., Russo A. G., Gervasi F., Braga M., and Renzi C., “Comorbidities, Socioeconomic Status, and Colorectal Cancer Diagnostic Route,” JAMA Network Open 8, no. 5 (2025): e258867, 10.1001/jamanetworkopen.2025.8867. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Black G., Sheringham J., Spencer‐Hughes V., et al., “Patients' Experiences of Cancer Diagnosis as a Result of an Emergency Presentation: A Qualitative Study,” PLoS One 10, no. 8 (2015): e0135027, 10.1371/journal.pone.0135027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Simmons L. H., May F. P., and Gilliam M. A., “Understanding the Older Patient Perspective on Colorectal Cancer Screening Cessation,” JAMA Network Open 7, no. 12 (2024): e2447806, 10.1001/jamanetworkopen.2024.47806. [DOI] [PubMed] [Google Scholar]
- 17. Sedani A. E., Gomez S. L., Lawrence W. R., et al., “Social Risks and Nonadherence to Recommended Cancer Screening Among US Adults,” JAMA Network Open 7, no. 1 (2024): e2145678. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Castañeda‐Avila M. A., Tisminetzky M., Oyinbo A. G., and Lapane K., “Racial and Ethnic Disparities in Use of Colorectal Cancer Screening Among Adults With Chronic Medical Conditions,” Preventing Chronic Disease 21 (2024): 230257, 10.5888/pcd21.230257. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Baeker, Bispo J., Islami F., Bandi P., Jemal A., and Most U. S., “Adults Overdue for Colorectal Cancer Screening Report Not Receiving Screening Recommendation From Doctor,” Annals of Internal Medicine 176 (2023): 1341. [DOI] [PubMed] [Google Scholar]
- 20. Sepassi A., Li M., Zell J. A., et al., “Rural‐Urban Disparities in Colorectal Cancer Screening, Diagnosis, Treatment, and Survivorship Care: A Systematic Review and Meta‐Analysis,” Oncologist 29, no. 4 (2024): e431–e446, 10.1093/oncolo/oyad347. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Castañeda‐Avila M. A., Tisminetzky M., Oyinbo A. G., and Lapane K. L., “Racial and Ethnic Disparities in Use of Colorectal Cancer Screening among Adults With Chronic Medical Conditions: BRFSS 2012–2020,” Preventing Chronic Disease 21 (2024): 230257, 10.5888/pcd21.230257. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Dillon E. C., Martinez M. C., Li M., et al., “It Is Not the Fault of the Health Care Team ‘It Is the Way the System Works’: A Mixed‐Methods Study of Fragmented Cancer Care,” BMC Health Services Research 24 (2024): 1378, 10.1186/s12913-024-11744-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Surveillance, Epidemiology, and End Results (SEER) Program. SEER Data Reporting Tools. Site Recode ICD‐O‐3/WHO 2008, accessed August 23, 2025. SEER Data.
- 24.Surveillance, Epidemiology, and End Results (SEER) Program. Available at: SEER Incidence Data, 1975–2021 (cancer.gov), accessed August 23, 2025.
- 25. National Cancer Institute . SEER‐Medicare: Medicare enrollment, claims, and assessment data, accessed August 24, 2025. healthservices.cancer.gov/seermedicare/medicare/.
- 26. National Cancer Institute . Overview of the SEER‐Medicare data, accessed August 24, 2025. jstor.org/stable/pdf/3767919.pdf.
- 27. Centers for Disease Control and Prevention ., CDC Social Vulnerability Index (SVI) [Internet] Atlanta (GA): Centers for Disease Control and Prevention, https://www.atsdr.cdc.gov/place. [Google Scholar]
- 28. Flanagan B. E., Gregory E. W., Hallisey E. J., et al., “A Social Vulnerability Index for Disaster Management,” Journal of Homeland Security and Emergency Management 8, no. 1 (2011): 1792, 10.2202/1547-7355.1792. [DOI] [Google Scholar]
- 29. de Graaff M. R., Elfrink A. K. E., and Buis C. I., “Defining Textbook Outcome in Liver Surgery and Assessment of Hospital Variation: A Nationwide Population‐Based Study,” European Journal of Surgical Oncology 48, no. 12 (2022): 2414–2423, 10.1016/j.ejso.2022.06.012. [DOI] [PubMed] [Google Scholar]
- 30. Steinzor P., Hernandez‐Boussard T., Doubeni C. A., et al., “Cost‐Effectiveness of Blood‐Based vs Stool‐Based Colorectal Cancer Screening in Underserved Populations: A Microsimulation Analysis,” JAMA Network Open 8, no. 1 (2025): e250112, 10.1001/jamanetworkopen.2025.0112. [DOI] [Google Scholar]
- 31. Mehta S. J., Pepe R. S., Polsky D., et al., “Cost‐Effectiveness of Noninvasive Colorectal Cancer Screening Strategies in Community Clinics,” JAMA Network Open 7, no. 2 (2024): e2339848, 10.1001/jamanetworkopen.2023.39848. [DOI] [Google Scholar]
- 32. O'Malley A. S., Forrest C. B., Feng S., and Mandelblatt J., “Disparities Despite Coverage: Gaps in Colorectal Cancer Screening Among Medicare Beneficiaries,” JAMA Internal Medicine 165, no. 18 (2005): 2129–2135, 10.1001/archinte.165.18.2129. [DOI] [PubMed] [Google Scholar]
- 33. Zimolzak A. J., Kapadia P., Upadhyay D. K., et al., “Frequent Missed Opportunities for Earlier Diagnosis of Advanced‐Stage Colorectal or Lung Cancer,” JAMA Internal Medicine 185, no. 9 (2025): 987–996, 10.1001/jamainternmed.2025.1837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Brenner H., Heisser T., and Hoffmeister M., “Delayed Cancer Registration and Estimation of Screening Colonoscopy Effects,” JAMA Network Open 7, no. 10 (2024): e2435669, 10.1001/jamanetworkopen.2024.35669. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Demb J., Kolb J. M., and Dounel J., “Red Flag Signs and Symptoms and Delays in Diagnosis for Early‐Onset Colorectal Cancer: A Systematic Review and Meta‐Analysis,” JAMA Network Open 7, no. 5 (2024): e2412345, 10.1001/jamanetworkopen.2024.12345. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Doubeni C. A., Simon M., and Krist A. H., “Addressing Systemic Racism Through Clinical Preventive Service Recommendations From the US Preventive Services Task Force,” Journal of the American Medical Association 325, no. 7 (2021): 627–628, 10.1001/jama.2020.26188. [DOI] [PubMed] [Google Scholar]
- 37. Sudore R. L. and Schillinger D., “Health Literacy and Systemic Racism—Using Clear Communication to Reduce Health Care Inequities,” JAMA Internal Medicine 183, no. 10 (2023): 1045–1046, 10.1001/jamainternmed.2023.3345. [DOI] [PubMed] [Google Scholar]
- 38. Metzl J. M., Maybank A., and De Maio F., “Responding to the COVID‐19 Pandemic: The Need for a Structurally Competent Health Care System,” Journal of the American Medical Association 324, no. 3 (2020): 231–232, 10.1001/jama.2020.14701. [DOI] [PubMed] [Google Scholar]
- 39. Wang C. and Shaukat A., “Optimal Approach to Colorectal Cancer Screening,” Gastroenterología y Hepatología 21, no. 3 (2025): e145–e152. [PMC free article] [PubMed] [Google Scholar]
- 40. Knudsen A. B., Rutter C. M., Peterse E. F., et al., “Colorectal Cancer Screening: An Updated Modeling Study for the US Preventive Services Task Force,” Journal of the American Medical Association 325, no. 19 (2021): 1998–2011, 10.1001/jama.2021.5746. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: Medical Procedure and Screening Codes Reference.
Table S2: Health and Financial Outcomes of Colon and Rectal Cancer Patients by Care Site.
Data Availability Statement
The data for this study were obtained from the linked SEER‐Medicare database. There are restrictions to the availability of this data, which is used under license for this study. Data can be accessed with permission from the National Cancer Institute and Center for Medicare and Medicaid Services.