Abstract
Purpose
Older Black women and women living in areas of low socioeconomic status (SES) diagnosed with cervical cancer (CC) have worse overall survival (OS). The objective was to investigate associations between OS and race/ethnicity and sociodemographic factors in younger (21–64 years) and older women (≥ 65 years) diagnosed with CC using Surveillance, Epidemiology, and End Results Program data.
Methods
This retrospective, population-based cohort study included 39,000 women ≥ 21 years diagnosed with CC diagnosed between 2006 and 2020. Age-group stratified Cox proportional hazards models adjusted for age, diagnosis year, and histology examined sociodemographic (rurality, SES, and persistent poverty) differences in OS.
Results
In the sample, 82.8% were < 65 years. Compared to younger women, older women were more likely to be non-Latinx (NL) Black (16.0 vs 12.9%) and diagnosed with late-stage CC (67.9 vs 47.5%). Adjusted models suggested younger NL Black women had worse OS than their NL White counterparts (HR 1.45 [95% CI 1.37–1.54]), this association was not found among older NL Black women (HR 1.06 [95% CI 0.96–1.16]). Similarly, younger women in lowest SES areas had worse OS compared to women in highest SES areas (HR 1.82 [95% CI 1.69–1.96]), this association was attenuated in older women (HR 1.27 [95% CI 1.15–1.42]). Finally, younger women living in persistent poverty had worse OS compared to those who did not (HR 1.40 [95% CI 1.32–1.48]), this association was not found in older women (HR 1.10 [95% CI 0.99–1.21]).
Conclusion
Sociodemographic disparities were found in CC OS for women < 65 that were attenuated or nonexistent in women ≥ 65 years.
Keywords: Cervical Cancer, Overall Survival, Older Populations, Surveillance, Epidemiology, End Results
Introduction
When diagnosed at an early-stage cervical cancer has over a 90% 5-year survival rate [1]. This overall survival rate changes based on factors such as stage at diagnosis, age, overall health, race/ethnicity, and sociodemographic factors (e.g. living in an area of poverty, insurance status, rurality, etc.) [2-4]. In the United States (US), identifying as Black or Latinx, living in rural areas or areas with low socioeconomic status (SES) have been associated with lower overall survival in people diagnosed with cervical cancer [4-10].
These previous studies tend to only include women younger than 65 years [8] or do not differentiate between younger and older populations in their analysis and simply adjust for age [4-7, 9, 10]. But, in the US, the accessibility, availability, and affordability of care change based on age [11]. In the US, a vast majority of women 65 years and older are enrolled in Medicare or Medicare Advantage plans and receive Social Security benefits [9, 12]. These benefits have been proven to improve the access, availability, and quality of care (e.g. access to a National Cancer Institute-accredited institute and/or access to a gynecologic oncologist) [13-16]. Younger women are also more likely to be uninsured or covered by Medicaid [17], which has been associated with higher rates of late-stage cervical cancer diagnosis and death [8, 18]. Finally, exposure to poverty is cumulative, meaning that individuals living for extended times in such areas are more likely to have lasting health effects [19, 20].
While cervical cancer is most often diagnosed in women aged 35–44, over 20% of the new cases of cervical cancer are diagnosed in women older than 65 years of age. Given the differences in access, availability, and quality of care based on age it is important to understand the association between sociodemographic factors (living in areas of low SES, areas of persistent poverty, and rural areas) on overall survival stratified for younger (< 65 years) and older (≥ 65 years) women diagnosed with cervical cancer.
Methods
Selection and description of study participants
The study utilized data from the Surveillance, Epidemiology, and End Results (SEER) Program Census Tract-level SES and Rurality Database [21] to conduct a retrospective, population-based cohort study on patients diagnosed with cervical cancer between 1 January, 2006 to 31 December, 2018 with follow up through 2020 (the most recent year available). The SEER program, which is funded by the National Institutes of Health National Cancer Institute, includes population-based cancer incidence data across 18 cancer registry regions and includes 48% of patients in the United States. Cancer diagnoses were identified using the International Classification of Diseases for Oncology, Version 3 (ICD-O-3) [18]. The specific database utilized for the study includes demographic information, survival information (including cause of death based on International Classification of Disease [ICD]-9 and −10), and linkage to census tract data from the American Community Survey to determine area-level rurality, SES, and persistent poverty based on patients’ residence at the time of cancer diagnosis [21].
Women aged 21 and older with a diagnosis of cervical cancer between 2006 and 2018 were included in our study (n = 39,000). Women who were diagnosed with cervical cancer at autopsy, per death certificate/data, and/or non-microscopically confirmed cases were excluded. In addition, patients diagnosed with multiple cancers and a histological subtype suggestive of another primary were excluded.
The University of Illinois, Chicago Institutional Review Board reviewed this study and waived approval and informed consent, determining that it did not involve human participant research.
Data collection
Demographic information (including linkage to area-level sociodemographic characteristics) for this study was collected at time of cervical cancer diagnosis from the SEER registry data. Race and ethnicity were coded as Latinx (all races), Non-Latinx (NL) American Indian/Alaskan Native (AIAN) (hereafter AIAN), NL Asian/Pacific Islander (API) (hereafter API), NL Black (hereafter Black) and NL White (hereafter White). Other variables collected included SEER summary diagnosis stage (localized, regional, or distant), treatment modalities received (surgery, chemotherapy, and/or radiation), histology (coded as squamous cell carcinoma, adenocarcinoma, adenosquamous carcinoma, neuroendocrine, and other). Overall survival was determined based on the number of months from the date of cervical cancer diagnosis (index time) to the earliest of date of death from any cause or date of last follow-up. Area-level sociodemographic characteristics were determined using the Census Tract-level SES and Rurality Databases. Rurality was defined using the US Department of Agriculture’s Rural–Urban Commuting Area (RUCA) 2010 census-level codes with a dichotomous classification: urban and rural [22]. SES was defined using census-based, US population-weighted Yost index (in quintiles), a time-dependent composite score derived from incorporates measures of poverty, occupation, education, and income [23]. The Yost quintiles were generated matching tumor diagnosis year with respective American Community Survey (ACS) 5-year estimates [21]. Finally, persistent poverty was defined (per the US Department of Agriculture) as Census tracts with a poverty rate 20% or more for at least five consecutive time periods, about 10 years apart, spanning approximately 30 years or more. SEER calculates these variables based on the 1990, 2000 decennial censuses, and 2007–2011 and 2015–2019 American Community Survey 5-year estimates [21, 24]. In our analytic cohort, women with unknown/missing rurality (n = 1,630), SES (n = 2,431), and/or persistent poverty (n = 1,630) indicators were excluded.
Statistical analysis
Demographic and clinical characteristics of women diagnosed with cervical cancer were compared by age (dichotomized into younger [< 65 years] and older [≥ 65 years] women) and race and ethnicity using descriptive statistics. To evaluate for differences between groups, the Wilcoxon rank-sum test was utilized for median values, the Z-test for proportions, and the Chi-square test for categorical variables (Fisher’s exact test for cells < 5). Overall survival was calculated using the Kaplan–Meier product-limit estimator with survival curves stratified by age group, race and ethnicity, rurality, SES, and persistent poverty were generated. Stratified Cox proportional hazard models (adjusted for clinical factors including age, diagnosis year, and histology per recommendations by the Institute of Medicine’s recommendations regarding disparities) [25-27] were used to estimate the hazard ratios (HR) of race and ethnicity on overall survival among younger (21–64 years) and older (≥ 65 years) women diagnosed with cervical cancer due to the known disparities in cervical cancer mortality for Black, Latinx, and American Indian/Alaskan Native women [5, 6, 28-30]. Stage at diagnosis was not adjusted for in our primary models as it has been shown to be a major mediator of overall survival [18] (a sensitivity analysis with stage included was also conducted for all models). Treatment related factors were also not adjusted for in our models as treatment was viewed as a mediator of overall survival rather than a confounding variable [25]. In addition, we also separately evaluated the association between each SES, rurality, and persistent poverty on overall survival in the stratified groups of younger and older women with cervical cancer to evaluate the effects of these sociodemographic factors on each age group specifically. We did not present fully adjusted models for all these sociodemographic factors in one model as they are each likely on the causal pathway and thus interrelated [31].
All p-values were from 2-sided tests and results were deemed statistically significant at p < 0.05. The proportional hazards assumption was evaluated using scaled Schoenfeld residuals, and no evidence suggesting non-proportionality was observed.
Results
This study included 39,000 women diagnosed with cervical cancer, of whom, 82.8% were < 65 years of age and 17.2% were ≥ 65 years of age. Among younger women (21–64 years of age), the median age was 45 (interquartile range [IQR] 37–53) years. Among older women (≥ 65 years of age), the median age was 72 (IQR 68–79). Among younger women, a higher proportion of women were identified as Latinx compared to older women (24.7 vs 18.4%, p < 0.01). In contrast, a higher proportion of older women were identified as Black compared to younger women (16.0 vs 12.9%, p < 0.01). Older women were much more likely to be diagnosed with late-stage cervical cancer (regional or distant SEER Summary stage) compared to younger women (67.9 vs 47.5%, p < 0.01). Additionally, older women were less likely to live in the lowest SES quintile (20.7 vs 22.1%, p < 0.01) and areas of persistent poverty (13.1 vs 11.9%, p = 0.04) compared to younger women. Similar proportions of younger and older women lived in rural areas (10.0 vs 10.3%, p = 0.74) (Table 1).
Table 1.
Descriptive characteristics of women diagnosed with cervical cancer by age group
| All women (n = 39,000) |
Ages < 65 years (n = 32,308) |
Ages 65 + years (n = 6,692) |
p | ||||
|---|---|---|---|---|---|---|---|
| n | (%) | n | (%) | n | (%) | ||
| Age, years | |||||||
| Median (IQR) | 48 | (39–60) | 45 | (37–53) | 72 | (68–79) | |
| Race/ethnicity | |||||||
| Non-Latinx White | 20,511 | (52.6) | 16,924 | (52.4) | 3,587 | (53.6) | < 0.01 |
| Non-Latinx Black | 5,247 | (13.5) | 4,175 | (12.9) | 1,072 | (16.0) | |
| Non-Latinx AIAN | 287 | (0.7) | 257 | (0.8) | 30 | (0.4) | |
| Non-Latinx API | 3,733 | (9.6) | 2,958 | (9.2) | 775 | (11.6) | |
| Latinx | 9,222 | (23.6) | 7,994 | (24.7) | 1,228 | (18.4) | |
| Year of diagnosis | |||||||
| 2006–2008 | 9,197 | (23.6) | 7,661 | (23.7) | 1,536 | (23.0) | 0.32 |
| 2009–2011 | 8,840 | (22.7) | 7,337 | (22.7) | 1,503 | (22.5) | |
| 2012–2014 | 8,768 | (22.5) | 7,266 | (22.5) | 1,502 | (22.4) | |
| 2015–2018 | 12,195 | (31.3) | 10,044 | (31.1) | 2,151 | (32.1) | |
| Histology | |||||||
| Squamous cell carcinoma | 26,515 | (68.0) | 21,735 | (67.3) | 4,780 | (71.4) | < 0.01 |
| Adenocarcinoma | 9,849 | (25.3) | 8,431 | (26.1) | 1,418 | (21.2) | |
| Adenosquamous | 1,409 | (3.6) | 1,226 | (3.8) | 183 | (2.7) | |
| Neuroendocrine | 628 | (1.6) | 507 | (1.6) | 121 | (1.8) | |
| Other | 599 | (1.5) | 409 | (1.3) | 190 | (2.8) | |
| Summary stage | |||||||
| Localized | 17,947 | (46.0) | 16,156 | (50.0) | 1,791 | (26.8) | < 0.01 |
| Regional | 14,367 | (36.8) | 11,266 | (34.9) | 3,101 | (46.3) | |
| Distant | 5,516 | (14.1) | 4,068 | (12.6) | 1,448 | (21.6) | |
| Unknown | 1,170 | (3.0) | 818 | (2.5) | 352 | (5.3) | |
| Time to treatment, months | |||||||
| Mean (SD) | 1.10 | (1.41) | 1.08 | (1.43) | 1.24 | (1.37) | < 0.01 |
| Surgery | |||||||
| No | 17,137 | (43.9) | 12,718 | (39.4) | 4,419 | (66.0) | < 0.01 |
| Yes | 21,863 | (56.1) | 19,590 | (60.6) | 2,273 | (34.0) | |
| Radiation | |||||||
| No | 17,765 | (45.6) | 15,377 | (47.6) | 2,388 | (35.7) | < 0.01 |
| Yes | 21,235 | (54.4) | 16,931 | (52.4) | 4,304 | (64.3) | |
| Chemotherapy | |||||||
| No | 20,014 | (51.3) | 16,658 | (51.6) | 3,356 | (50.1) | 0.04 |
| Yes | 18,986 | (48.7) | 15,650 | (48.4) | 3,336 | (49.9) | |
| Any systemic treatment | |||||||
| No | 2,652 | (6.8) | 1,668 | (5.2) | 984 | (14.7) | < 0.01 |
| Yes | 36,348 | (93.2) | 30,640 | (94.8) | 5,708 | (85.3) | |
| Marital status | |||||||
| Not married | 20,372 | (52.2) | 16,174 | (50.1) | 4,198 | (62.7) | < 0.01 |
| Married | 16,197 | (41.5) | 14,172 | (43.9) | 2,025 | (30.3) | |
| Unknown | 2,431 | (6.2) | 1,962 | (6.1) | 469 | (7.0) | |
| Yost US quintile | |||||||
| Quintile 1—lowest SES | 8,533 | (21.9) | 7,147 | (22.1) | 1,386 | (20.7) | < 0.01 |
| Quintile 2 | 7,408 | (19.0) | 6,183 | (19.1) | 1,225 | (18.3) | |
| Quintile 3 | 6,828 | (17.5) | 5,687 | (17.6) | 1,141 | (17.1) | |
| Quintile 4 | 7,057 | (18.1) | 5,798 | (17.9) | 1,259 | (18.8) | |
| Quintile 5—highest SES | 7,040 | (18.1) | 5,738 | (17.8) | 1,302 | (19.5) | |
| Unknown | 2,134 | (5.5) | 1,755 | (5.4) | 379 | (5.7) | |
| RUCA category | |||||||
| Urban | 33,459 | (85.8) | 27,738 | (85.9) | 5,721 | (85.5) | 0.74 |
| Rural | 3,911 | (10.0) | 3,225 | (10.0) | 686 | (10.3) | |
| Unknown | 1,630 | (4.2) | 1,345 | (4.2) | 285 | (4.3) | |
| Poverty | |||||||
| Not persistent poverty | 32,355 | (83.0) | 26,746 | (82.8) | 5,609 | (83.8) | 0.04 |
| Persistent poverty | 5,015 | (12.9) | 4,217 | (13.1) | 798 | (11.9) | |
| Unknown | 1,630 | (4.2) | 1,345 | (4.2) | 285 | (4.3) | |
When stratified by race/ethnicity, Black women were more likely to be diagnosed with late-stage cervical cancer compared to White women (58.6 vs 49.4%, p < 0.01), more likely to live in the lowest SES quintile (46.7 vs 14.8%, p < 0.01) and in areas of persistent poverty compared to White women (31.4 vs 6.5%, p < 0.01). Further demographics stratified by race/ethnicity are available in Supplemental Table 1.
Among younger women, 5-year overall survival was 70.8%. There were also substantial sociodemographic inequities in overall survival (Fig. 1 and Table 2). Black women had worse overall survival than White women (median survival: 35 vs. 48 months, HR: 1.45, 95% confidence interval [CI] 1.37–1.54). In contrast, younger API and Latinx women had improved overall survival compared to White women (HR: 0.86, 95% CI 0.80–0.94 and HR: 0.94, 95% CI 0.89–0.99; respectively). Younger Women in rural areas also had worse overall survival than women living in urban areas (median survival time 42 vs. 43 months, HR: 1.13, 95%CI 1.05–1.21). Younger women living in areas with the lowest SES also had worse overall survival than women living in the areas of highest SES (median survival: 36 vs 53 months, HR: 1.82, 95% CI 1.6–1.96). When investigating younger women living in persistent poverty had a worse overall survival than women not living in persistent poverty (median survival time 35 vs 44 months, HR: 1.40, 95% CI 1.32–1.48). Finally, when considering stage, younger women diagnosed with advanced-stage cervical cancer had worse overall survival than those diagnosed with early-stage cervical cancer (median survival time 59 months vs not reached due to high overall survival, HR: 5.29, 95% CI 5.03–5.56) (Fig. 1 and Table 2).
Fig. 1.
Kaplan–Meier survivor functions for overall survival stratified by age (< 65/≥ 65 years) for race/ethnicity, Yost Quintile, RUCA category, persistent poverty. Abbreviations: NL, Non-Latinx; AI/AN, American Indian/Alaskan Native; API, Asian/Pacific Islander
Table 2.
Associations between race/ethnicity and area-level social determinants of health with overall survival among women diagnosed with cervical cancer
| Unadjusted model |
Multivariable model* |
|||||
|---|---|---|---|---|---|---|
| HR | (95% CI) | P | HR | (95% CI) | P | |
| Ages < 65 years | ||||||
| Race/ethnicity | ||||||
| Non-latinx white | Reference | Reference | ||||
| Non-latinx black | 1.59 | (1.50–1.68) | < 0.01 | 1.45 | (1.37–1.54) | < 0.01 |
| Non-latinx AIAN | 1.09 | (0.86–1.39) | 0.46 | 1.14 | (0.90–1.44) | 0.29 |
| Non-latinx API | 0.93 | (0.86–1.01) | 0.08 | 0.86 | (0.80–0.94) | < 0.01 |
| Latinx | 0.91 | (0.86–0.96) | < 0.01 | 0.94 | (0.89–0.99) | 0.02 |
| Yost US quintile | ||||||
| Quintile 1—lowest SES | 1.91 | (1.78–2.06) | < 0.01 | 1.82 | (1.69–1.96) | < 0.01 |
| Quintile 2 | 1.59 | (1.48–1.72) | < 0.01 | 1.55 | (1.43–1.67) | < 0.01 |
| Quintile 3 | 1.44 | (1.33–1.55) | < 0.01 | 1.41 | (1.30–1.52) | < 0.01 |
| Quintile 4 | 1.22 | (1.13–1.32) | < 0.01 | 1.21 | (1.12–1.31) | < 0.01 |
| Quintile 5—highest SES | Reference | Reference | ||||
| RUCA category | ||||||
| Urban | Reference | Reference | ||||
| Rural | 1.15 | (1.07–1.23) | < 0.01 | 1.13 | (1.05–1.21) | < 0.01 |
| Poverty | ||||||
| Not persistent poverty | Reference | Reference | ||||
| Persistent poverty | 1.45 | (1.37–1.53) | < 0.01 | 1.40 | (1.32–1.48) | < 0.01 |
| Stage | ||||||
| Local | Reference | Reference | ||||
| Regional/distant | 7.27 | (6.87, 7.71) | 5.29 | (5.03, 5.56) | < 0.01 | |
| Ages 65 + years | ||||||
| Race/ethnicity | ||||||
| Non-atinx white | Reference | Reference | ||||
| Non-latinx black | 1.12 | (1.02–1.22) | 0.02 | 1.06 | (0.96–1.16) | 0.25 |
| Non-latinx AIAN | 1.69 | (1.09–2.62) | 0.02 | 1.74 | (1.12–2.71) | 0.01 |
| Non-latinx API | 0.70 | (0.63–0.79) | < 0.01 | 0.67 | (0.60–0.76) | < 0.01 |
| Latinx | 0.82 | (0.75–0.90) | < 0.01 | 0.83 | (0.75–0.91) | < 0.01 |
| Yost US quintile | ||||||
| Quintile 1—lowest SES | 1.24 | (1.12–1.39) | < 0.01 | 1.27 | (1.15–1.42) | < 0.01 |
| Quintile 2 | 1.16 | ()1.04–1.29 | < 0.01 | 1.19 | (1.07–1.33) | < 0.01 |
| Quintile 3 | 1.12 | (1.00–1.25) | 0.05 | 1.11 | (1.00–1.25) | 0.06 |
| Quintile 4 | 1.10 | (0.99–1.23) | 0.08 | 1.10 | (0.99–1.23) | 0.09 |
| Quintile 5—highest SES | Reference | Reference | ||||
| RUCA category | ||||||
| Urban | Reference | Reference | ||||
| Rural | 1.08 | (0.97–1.20) | 0.19 | 1.11 | (1.00–1.24) | 0.06 |
| Poverty | ||||||
| Not persistent poverty | Reference | Reference | ||||
| Persistent poverty | 1.09 | (0.98–1.20) | 0.10 | 1.10 | (0.99–1.21) | 0.07 |
| Stage | ||||||
| Local | Reference | Reference | ||||
| Regional/Distant | 2.93 | (2.67, 3.21) | < 0.01 | 2.85 | (2.61, 3.13) | < 0.01 |
Multivariable estimates are adjusted in separate models for age (continuous), diagnosis year (categorical; 2006–2008, 2009–2011, 2012–2014, 2015–2018) and histology (categorical; squamous cell carcinoma, adenocarcinoma, adenosquamous, neuroendocrine, other)
In contrast, women older than 65 years of age had a lower overall survival (5-year overall survival was 42.6%), but the magnitude of sociodemographic disparities in overall survival were attenuated or had wider confidence intervals and were non-significant (Fig. 1 and Table 2). Older Black women had similar overall survival compared with older White women (median survival time: 18 vs. 23 months, HR: 1.06, 95% CI 0.96–1.16). Additionally, older API and Latinx had similar overall survival compared to their White counterparts as their younger counterparts, (HR: 0.86, 95% CI 0.80–0.94 and HR: 0.94, 95% CI 0.89–0.99, respectively). Older Women living in the lowest SES quintile had a worse overall survival compared to women living in the highest SES quintile (median survival time: 20.5 vs. 24 months, HR 1.27, 95% CI 1.15–1.42), although the difference in overall survival was attenuated compared to the difference seen in patients < 65 years of age. In addition, there were no statistically significant difference in overall survival in women ≥ 65 years living in rural areas compared to urban areas or areas of persistent poverty compared to areas without persistent poverty (median survival time: 22 vs. 22 months, HR: 1.11, 95% CI 1.00–1.24; median survival time: 23 vs. 21 months, HR: 1.10, 95%CI 0.99–1.21, respectively). When evaluating overall survival by stage in older women, older women diagnosed with advanced-stage cervical cancer had worse overall survival than older women diagnosed with early-stage cervical cancer (median survival time: 25 vs 130 months, HR: 2.85, 95% CI 2.61–3.13).
Finally, in our sensitivity analysis (which additionally included adjustment for stage at diagnosis) evaluating association between race/ethnicity and area-level social determinants of health with overall survival among women diagnosed with cervical cancer found similar associations between race/ethnicity and area-level social determinants of health with overall survival among younger and older women diagnosed with cervical cancer. Most importantly, the sensitivity analysis also showed similar attenuations and lack of associations as the primary analysis between older Black women, women living in the lowest SES quintile, and women living in persistent poverty diagnosed with cervical cancer (Supplemental Table 2).
Discussion
In this study of 39,000 women diagnosed with cervical cancer in SEER areas from 2006 to 2018, we found that women over the age of 65 had higher rates of late-stage cancer and lower rates of 5-year survival compared to their younger counterparts, consistent with previous findings [32, 33]. However, our study is the first to characterize the association with race/ethnicity and sociodemographic factors in women diagnosed with cervical cancer over the age of 65 years and respective overall survival.
In our younger cohort of women diagnosed with cervical cancer under the age of 65 we found results consistent with the previous literature [4-6, 9, 10, 33]. Black women compared to White women, women living in areas of lower SES compared to the highest SES, women living in rural areas compared to urban areas, and women living in persistent poverty compared to those who did not have much worse overall survival than their respective reference groups. We expected that these relative differences in overall survival would significantly widen in aging vulnerable populations, as exposure to poverty and social determinants of health can be cumulative and exacerbate poor outcomes, exposing older populations to more competing causes of death [19, 20]. Rather we found the opposite, that the relative difference between the respective groups of older women and their comparison groups were significantly attenuated or not significant at all (even after adjustment for stage in our sensitivity analysis). These findings expand on a previous study that investigated outcomes in older women diagnosed with cervical cancer in the Medicare-SEER database that found no association between race/ethnicity and cancer-specific survival [34].
These findings of differential overall survival between older women diagnosed with cervical cancer and younger women could be due to the benefits of Medicare and Social Security, which are obtainable after the age of 65 in the United States [14, 35-37]. Medicare and Social Security provide individuals who have met federal work requirements, with healthcare benefits and a basic income. These benefits are nearly universal, with 90% of adults older than 65 years and older receiving Social Security [12] and 99% receiving Medicare [38]. Medicare eligibility has long been proven to be critical in improving healthcare access for low income women [39] and women living in rural areas [40]. For gynecological cancers, in particular, enrollees in Medicare plans generally have better access to gynecologist oncologists, NCI-designated cancer center, clinical trials, and more timely access to care after cancer diagnosis [16, 41]. In contrast, among women ages 19–64, 10.8% are uninsured and 15.8% are covered by Medicaid [42]. While our study does not include insurance as a variable (due to the nature of SEER data availability), these findings may speak to the potential benefit Medicare provides compared to Medicaid. Previous studies have found that being uninsured or covered by Medicaid mediated over 50% of the advanced-stage cervical cancers diagnosed and worsened cervical cancer specific survival in minoritized populations compared to those covered by private insurance/Medicare [8, 18]. This could be due to the decrease in access to gynecologic oncologists, guideline-concordant cancer care, NCI-designated cancer center, affordable care, clinical trials, and increased timely treatment after cancer diagnosis [16, 41].
These findings potentially showcase the benefits these social programs provide our vulnerable adults living in rural areas and areas of poverty. This is especially relevant as older Black and Latinx women are more likely to be living below the federal poverty line than younger Black and Latinx women per the most recent US census data [43]. Social Security is one of the most important anti-poverty programs in the United States and continues to improve the health of our nation’s most vulnerable populations [14, 44]. It is important to note though, that our study only found that the association between lowest SES and highest SES and overall survival was attenuated implying that, if our hypothesis is correct, more than insurance coverage and a guaranteed income is needed to alleviate this disparity. Future policy changes should consider expansion of these benefits for vulnerable populations and carefully evaluate both the potential benefits and harms related to the expansion [45, 46]. Additionally, Implementation of a basic income for younger vulnerable populations and improvements in the Medicaid insurance coverage (e.g., mandating that coverage include a gynecologic oncologist) could potentially begin to ameliorate the disparities seen in cervical cancer overall survival and potentially other preventable diseases.
While our study may be explained by the insurance and social benefits that are only available to older women, it could be explained by a host of other issues that tie these groups of older women together. For instance, our study found that older women were more likely to be diagnosed with advanced-stage cervical cancer than younger women which may lead to less differences in overall survival between groups. These increased rates of advanced-stage cancer could potentially be due to poor access to care prior to turning 65 and being eligible for Medicare. Previous studies have found that utilization of healthcare services increase after individuals become Medicare eligible [15]. This could especially be critical as many women prior to the age 65 are not adequately screened per the current national guidelines and recommendations [47, 48]. Previous studies have also found that many providers over rely on self-reported screening history to end cervical cancer screening [49]. Thus, it is critical that healthcare providers carefully review their patients’ screening history, especially in older populations that meet the age requirement to end cervical cancer screening. Additionally, these older women, regardless of race and sociodemographic factors, may have multiple comorbidities, increased frailty, and a limited life expectancy across all groups that may preclude them from receiving treatment, which our study found a large proportion of older women did not receive any form of treatment compared to younger women. This could be why the overall survival for older women with early-stage cancer is much less than younger women diagnosed with early-stage cervical cancer. Future studies could investigate the effects of comorbidities and frailty on sociodemographic factors and cervical cancer overall survival to grasp a better understanding of disparities.
It is important to note that while we found younger Black women had worse overall survival than White women, this does not imply that race is a biological construct, rather it is a social construct [50-52]. Race is a broad proxy for the lived experience, shared, culture, and structural racism that is experienced by marginalized groups, despite this, race and ethnicity were explored due to the known disparities in cervical cancer. For instance, in our study, Black women had the highest proportions diagnosed with late-stage cervical cancer, living in areas with the lowest SES, and in areas of persistent poverty, which could be why these women also had the lowest overall survival. Future studies could investigate the mediation effect of these sociodemographic factors on cervical cancer survival inequities in younger and older Black women and other racial and ethnic groups.
This study has several limitations. First, SEER may not be generalizable to the entire population of the US [53]. Second, if a patient moved during the study period to an area not covered by the SEER registry region, they may be lost to follow up and may bias the study [54]. Third, our study did not directly evaluate the role of insurance status as it is no longer a variable that is actively updated in the SEER dataset [55]. Future studies investigating inequities in cervical cancer survival between older and younger women that includes insurance may better elucidate the possible relationship between age and overall survival described in this study. Fourth, our study relies on race and ethnicity data which are coded using the underlying cancer registry, depending on how the cancer registry identifies race and ethnicity there was been a history of misclassification compared to self-reported data [56, 57].Fifth, while SEER includes 18 states and approximately 48% of the US population, it may underrepresent rural areas and certain rural area populations [58]. Though the RUCA definition is able to provide an adequate estimate on differences between rural–urban disparities [59]. Sixth, individual-level data are not available in this SEER dataset, thus our models could not consider individual level socioeconomic factors or clinical factors such as comorbidities which may influence associations. Finally, our sample size for populations ≥ 65 years is relatively smaller compared to those < 65 years, leading to wider confidence intervals in these groups. Despite this, the hazard ratios and confidence intervals still demonstrate differences between the overall survival in younger and older populations when evaluating by race/ethnicity and the different sociodemographic characteristics.
Our study found that younger Black women and women facing significant sociodemographic inequities diagnosed with cervical cancer had worse overall survival than their White and more well-off counterparts. These associations were significantly attenuated or nonsignificant in these groups of women 65 years and older diagnosed with cervical cancer. Efforts should be made to ensure equitable access to care regardless of age.
Supplementary Material
Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s10552-025-01961-0.
Funding
Hunter Holt was supported by the University of Illinois at Chicago (UIC)’s Building Interdisciplinary Research Careers in Women's Health (BIRCWH) Grant K12HD101373 from the National Institutes of Health (NIH) Office of Research on Women's Health (ORWH). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
Ethical approval The University of Illinois Chicago Institutional Review Board reviewed this study and waived approval and informed consent, determining that it did not involve human participant research.
Data availability
Greg Calip 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. Data derived from a source in the public domain: Surveillance, Epidemiology, and End Results Program data. More information available at: https://seer.cancer.gov/data/specialized/available-databases/census-tract-request/
References
- 1.Cancer Stat Facts: Cervical Cancer (2022) Available from: https://seer.cancer.gov/statfacts/html/cervix.html.
- 2.Afshar N, English DR, Milne RL (2021) Factors explaining socioeconomic inequalities in cancer survival: a systematic review. Cancer Control 28:107327482110119 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Clark TG et al. (2003) Survival analysis Part I: basic concepts and first analyses. Br J Cancer 89(2):232–238 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kucera CW et al. (2024) Survival disparities in non-Hispanic Black and White cervical cancer patients vary by histology and are largely explained by modifiable factors. Gynecol Oncol 184:224–235 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Brookfield KF et al. (2009) Disparities in survival among women with invasive cervical cancer. Cancer 115(1):166–178 [DOI] [PubMed] [Google Scholar]
- 6.Singh GK (2012) Rural-urban trends and patterns in cervical cancer mortality, incidence, stage, and survival in the United States, 1950–2008. J Community Health 37(1):217–223 [DOI] [PubMed] [Google Scholar]
- 7.Singh GK, Jemal A (2017) Socioeconomic and racial/ethnic disparities in cancer mortality, incidence, and survival in the United States, 1950–2014: over six decades of changing patterns and widening inequalities. J Environ Public Health 2017:1–19 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Markt SC et al. (2018) Insurance status and cancer treatment mediate the association between race/ethnicity and cervical cancer survival. PLoS ONE 13(2):e0193047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Cohen CM et al. (2023) Racial and ethnic disparities in cervical cancer incidence, survival, and mortality by histologic subtype. J Clin Oncol 41(5):1059–1068 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Eggleston KS et al. (2006) Cervical cancer survival by socioeconomic status, race/ethnicity, and place of residence in Texas, 1995–2001. J Womens Health 15(8):941–951 [DOI] [PubMed] [Google Scholar]
- 11.Jacobs PD (2021) The impact of medicare on access to and affordability of health care. Health Aff 40(2):266–273 [DOI] [PubMed] [Google Scholar]
- 12.Administration, U.S.S.S., FACT SHEET: Social Security. 2023.
- 13.Medicare’s Current And Future Role In Reducing Racial And Ethnic Health Disparities. 2020. March 23, 2020. Available from: https://www.healthaffairs.org/content/forefront/medicare-s-current-and-future-role-reducing-racial-and-ethnic-health-disparities.
- 14.Arno PS et al. (2011) Social security and mortality: The role of income support policies and population health in the United States. J Public Health Policy 32(2):234–250 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Bartelt K, Haddock J, Piff A, Sahakian S, Franklin B, Allen S, Bryer E, McNitt J (2023) The ‘Medicare Effect’: Screening and Diagnosis Rates Increase in the First Year of Medicare Coverage. Available from: https://epicresearch.org/articles/the-medicare-effect-screening-and-diagnosis-rates-increase-in-the-first-year-of-medicare-coverage. [Google Scholar]
- 16.Bodurtha Smith AJ, Pena D, Ko E (2022) Insurance-mediated disparities in gynecologic oncology care. Obstet Gynecol 139(2):305–312 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Conway D (2020) Adults age 26 had highest uninsured rate among all ages, followed by 27-year-olds. 2020 October 8, 2021 [cited 2024 April 1]. Available from: https://www.census.gov/library/stories/2020/10/uninsured-rates-highest-for-young-adults-aged-19-to-34.html. [Google Scholar]
- 18.Holt HK et al. (2023) Mediation of racial and ethnic inequities in the diagnosis of advanced-stage cervical cancer by insurance status. JAMA Netw Open 6(3):e232985. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Hedman L et al. (2013) Cumulative exposure to disadvantage and the intergenerational transmission of neighbourhood effects. J Econ Geogr 15(1):195–215 [Google Scholar]
- 20.Kasper JD et al. (2008) Effects of poverty and family stress over three decades on the functional status of Older African American Women. J Gerontol B Psychol Sci Soc Sci 63(4):S201–S210 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Census Tract-level SES and Rurality Database (2006–2020). [cited 2022 Jul 14]. Available from: https://seer.cancer.gov/seerstat/databases/census-tract/index.html.
- 22.United States Department of Agriculture. Rura-urban commuting area codes. 2023. September 25, 2023 [cited 2023 October 19]. Available from: https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes/.
- 23.Yost K et al. (2001) Cancer Causes Control 12(8):703–711 [DOI] [PubMed] [Google Scholar]
- 24.US DEPARTMENT OF AGRICULTURE (2023) Poverty Area Measures. 2023 August 8, 2023 [cited 2023 October 19]. Available from: https://www.ers.usda.gov/data-products/poverty-area-measures/.
- 25.Cook BL, McGuire TG, Zaslavsky AM (2012) Measuring racial/ethnic disparities in health care: methods and practical issues. Health Serv Res 47(3 Pt 2):1232–1254 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.McGuire TG et al. (2006) Implementing the Institute of Medicine definition of disparities: an application to mental health care. Health Serv Res 41(5):1979–2005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Nelson A (2002) Unequal treatment: confronting racial and ethnic disparities in health care. J Natl Med Assoc 94(8):666. [PMC free article] [PubMed] [Google Scholar]
- 28.Beavis AL, Gravitt PE, Rositch AF (2017) Hysterectomy-corrected cervical cancer mortality rates reveal a larger racial disparity in the United States. Cancer 123(6):1044–1050 [DOI] [PubMed] [Google Scholar]
- 29.Islami F, Fedewa SA, Jemal A (2019) Trends in cervical cancer incidence rates by age, race/ethnicity, histological subtype, and stage at diagnosis in the United States. Prev Med 123:316–323 [DOI] [PubMed] [Google Scholar]
- 30.Yoo W et al. (2017) Recent trends in racial and regional disparities in cervical cancer incidence and mortality in United States. PLoS ONE 12(2):e0172548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Westreich D, Greenland S (2013) The Table 2 fallacy: presenting and interpreting confounder and modifier coefficients. Am J Epidemiol 177(4):292–298 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.SEER*Explorer: An interactive website for SEER cancer statistics. [cited 2023 May 10]. Available from: https://seer.cancer.gov/statistics-network/explorer/.
- 33.Cooley JJP et al. (2023) Cervical cancer stage at diagnosis and survival among women ≥65 years in California. Cancer Epidemiol Biomark Prev 32(1):91–97 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Lichter KE et al. (2022) Understanding cervical cancer after the age of routine screening: characteristics of cases, treatment, and survival in the United States. Gynecol Oncol 165(1):67–74 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Goldman DP et al. (2021) Improved survival for individuals with common chronic conditions in the Medicare population. Health Econ 30(S1):80–91 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Silvestri GA et al. (2021) Cancer outcomes among medicare beneficiaries and their younger uninsured counterparts: study compares cancer outcomes among Medicare beneficiaries and their younger uninsured counterparts. Health Aff 40(5):754–762 [DOI] [PubMed] [Google Scholar]
- 37.Wallace J et al. (2021) Changes in racial and ethnic disparities in access to care and health among US Adults at age 65 years. JAMA Intern Med 181(9):1207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Cohen RA et al. (2021) Demographic variation in health insurance coverage: United States, 2019. Natl Health Stat Report 159:1–15 [PubMed] [Google Scholar]
- 39.Aggarwal R et al. (2022) Medicare eligibility and healthcare access, affordability, and financial strain for low- and higher-income adults in the United States: a regression discontinuity analysis. PLoS Med 19(10):e1004083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Stearns SC, Slifkin RT, Edin HM (2000) Access to care for rural medicare beneficiaries. J Rural Health 16(1):31–42 [DOI] [PubMed] [Google Scholar]
- 41.Courtney-Brooks M et al. (2013) The Patient Protection and Affordable Care Act: impact on the care of gynecologic oncology patients in the absence of Medicaid expansion in central Virginia. Gynecol Oncol 130(2):346–349 [DOI] [PubMed] [Google Scholar]
- 42.Katherine Keisler-Starkey LNB, Rachel AL (2023) Health Insurance Coverage in the United States: 2022. 2023, United States Census Bureau. [Google Scholar]
- 43.Shrider EA, Burns K, Chen F, Semega J (2021) Poverty status of people by age, race, and Hispanic origin: 1959 to 2020. 2021, United States Cenus Bureau. [Google Scholar]
- 44.Amy Matsui CA (2023) Poverty data shows why social security matters to women and people of color. 2023 Dec 7, 2023 [cited 2023 Dec 14]. Available from: https://blog.ssa.gov/poverty-data-shows-why-social-security-matters-to-women-and-people-of-color/#:~:text=Social%20Security%20continued%20to%20be,of%20SPM%20poverty%20in%202022. [Google Scholar]
- 45.Altman NJ, Kingson ER (2021) Expanding social security to address the retirement income crisis. Public Policy Aging Report 31(3):102–107 [Google Scholar]
- 46.Holahan J, Buettgens M, Green A, Simpson M, Banthin J. Lowering the Age of Medicare Eligibility to 60: Effects on Coverage and Spending [Google Scholar]
- 47.Alimena S et al. (2024) Accessibility of criteria to exit cervical cancer screening at age 65 years in the electronic health record. O&G Open 1(3):032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Cejtin HE, Schmidt JB (2020) Prevalence of inadequate cervical cancer screening in low-income older women. J Womens Health 29(10):1350–1353 [DOI] [PubMed] [Google Scholar]
- 49.Holt HK et al. (2023) A qualitative study of primary care clinician’s approach to ending cervical cancer screening in older women in the United States. Preventive Medicine Reports 36:102500. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Braveman P, Parker Dominguez T (2021) Abandon "Race." Focus on racism. Front Public Health 9:689462. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Caulfield T et al. (2009) Race and ancestry in biomedical research: exploring the challenges. Genome Med 1(1):8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Roberts DE (2008) Is race-based medicine good for us?: African American approaches to race, biomedicine, and equality. J Law Med Ethics 36(3):537–545 [DOI] [PubMed] [Google Scholar]
- 53.Kuo T-M, Mobley LR (2016) How generalizable are the SEER registries to the cancer populations of the USA? Cancer Causes Control 27(9):1117–1126 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Yu JB et al. (2009) NCI SEER public-use data: applications and limitations in oncology research. Oncology (Williston Park) 23(3):288–295 [PubMed] [Google Scholar]
- 55.SEER. Insurance Recode (2007+). [cited 2022 Jul 14]. Available from: https://seer.cancer.gov/seerstat/variables/seer/insurance-recode/.
- 56.Clarke LC et al. (2016) Validity of race, ethnicity, and national origin in population-based cancer registries and rapid case ascertainment enhanced with a Spanish surname list. Med Care 54(1):e1–e8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57.Atekruse SF et al. (2017) Comparing cancer registry abstracted and self-reported data on race and ethnicity. J Registry Manag 44(1):30–33 [PubMed] [Google Scholar]
- 58.Zahnd WE et al. (2019) Challenges of using nationally representative, population-based surveys to assess rural cancer disparities. Prev Med 129:105812. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Long JC, Delamater PL, Holmes GM (2021) Which definition of rurality should I use?: The relative performance of 8 federal rural definitions in identifying rural-urban disparities. Med Care 59(10 Suppl 5):S413–S419 [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
Data Availability Statement
Greg Calip 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. Data derived from a source in the public domain: Surveillance, Epidemiology, and End Results Program data. More information available at: https://seer.cancer.gov/data/specialized/available-databases/census-tract-request/