Neighborhood socioeconomic deprivation, tumor subtypes, and causes of death after non-metastatic invasive breast cancer diagnosis: a multilevel competing-risk analysis

The Role of Residential Segregation in Treatment and Outcomes of Ductal Carcinoma In Situ of the Breast

BACKGROUND

It remains unclear whether residential segregation impacts on clinical treatment and outcomes for ductal carcinoma in situ (DCIS), a nonobligate precursor to invasive breast cancer (IBC).

METHODS

This population-based retrospective cohort study included adult non-Hispanic White and Black women diagnosed with unilateral DCIS between January 1990 and December 2015, followed through December 2016, and identified from the Surveillance, Epidemiology, and End Results dataset. County-level racialized economic segregation was measured using the Index of Concentration at the Extremes. Multilevel logistic regression and Cox proportional hazards regression accounting for county-level clustering were used to estimate the ORs of local treatment and HRs of subsequent IBC and mortality.

RESULTS

Of 103,898 cases, mean age was 59.5 years, 12.5% were non-Hispanic Black, 87.5% were non-Hispanic White, 97.5% underwent surgery, 64.5% received radiotherapy following breast-conserving surgery (BCS), 7.1% developed IBC, and 18.6% died from all causes. Among women living in the least versus most privileged counties, we observed higher odds of receiving mastectomy [vs. BCS; OR = 1.51; 95% confidence interval (CI), 1.35-1.69; Ptrend < 0.001] and radiotherapy following BCS(OR = 1.27; 95% CI, 1.07-1.51; Ptrend < 0.01); the risk was higher in subsequent ipsilateral IBC (HR = 1.16; 95% CI, 1.02-1.32; Ptrend = 0.04), not in breast cancer-specific mortality (HR = 1.04; 95% CI, 0.88-1.23; Ptrend = 0.56).

CONCLUSIONS

The results provide evidence for disparities in clinical treatment for DCIS and prognostic outcomes among women in racially and economically segregated counties.

IMPACT

Our findings may inform geographically targeted multilevel interventions to reduce breast cancer burden and improve breast cancer care and equity.

Neighborhood Deprivation and Breast Cancer Mortality Among Black and White Women

Importance

Neighborhood deprivation has been associated with increased breast cancer mortality among White women, but findings are inconsistent among Black women, who experience different neighborhood contexts. Accounting for interactions among neighborhood deprivation, race, and other neighborhood characteristics may enhance understanding of the association.

Objective

To investigate whether neighborhood deprivation is associated with breast cancer mortality among Black and White women and whether interactions with rurality, residential mobility, and racial composition, which are markers of access, social cohesion, and segregation, respectively, modify the association.

Design, Setting, and Participants

This population-based cohort study used Georgia Cancer Registry (GCR) data on women with breast cancer diagnosed in 2010 to 2017 and followed-up until December 31, 2022. Data were analyzed between January 2023 and October 2023. The study included non-Hispanic Black and White women with invasive early-stage (I-IIIA) breast cancer diagnosed between 2010 and 2017 and identified through the GCR.

Exposures

The Neighborhood Deprivation Index (NDI), assessed in quintiles, was derived through principal component analysis of 2011 to 2015 block group-level American Community Survey (ACS) data. Rurality, neighborhood residential mobility, and racial composition were measured using Georgia Public Health Department or ACS data.

Main Outcomes and Measures

The primary outcome was breast cancer-specific mortality identified by the GCR through linkage to the Georgia vital statistics registry and National Death Index. Cox proportional hazards regression was used to estimate age-adjusted and multivariable-adjusted hazard ratios (HRs) and 95% CIs for the association between neighborhood deprivation and breast cancer mortality.

Results

Among the 36 795 patients with breast cancer (mean [SD] age at diagnosis, 60.3 [13.1] years), 11 044 (30.0%) were non-Hispanic Black, and 25 751 (70.0%) were non-Hispanic White. During follow-up, 2942 breast cancer deaths occurred (1214 [41.3%] non-Hispanic Black women; 1728 [58.7%] non-Hispanic White women). NDI was associated with an increase in breast cancer mortality (quintile 5 vs 1, HR, 1.36; 95% CI, 1.19-1.55) in Cox proportional hazards models. The association was present only among non-Hispanic White women (quintile 5 vs 1, HR, 1.47; 95% CI, 1.21-1.79). Similar race-specific patterns were observed in jointly stratified analyses, such that NDI was associated with increased breast cancer mortality among non-Hispanic White women, but not non-Hispanic Black women, irrespective of the additional neighborhood characteristics considered.

Conclusions and Relevance

In this cohort study, neighborhood deprivation was associated with increased breast cancer mortality among non-Hispanic White women. Neighborhood racial composition, residential mobility, and rurality did not explain the lack of association among non-Hispanic Black women, suggesting that factors beyond those explored here may contribute to breast cancer mortality in this racial group.

Preexisting Diabetes and Breast Cancer Treatment Among Low-Income Women

IMPORTANCE

Diabetes is associated with poorer prognosis of patients with breast cancer. The association between diabetes and adjuvant therapies for breast cancer remains uncertain.

OBJECTIVE

To comprehensively examine the associations of preexisting diabetes with radiotherapy, chemotherapy, and endocrine therapy in low-income women with breast cancer.

DESIGN, SETTING, AND PARTICIPANTS

This population-based cohort study included women younger than 65 years diagnosed with nonmetastatic breast cancer from 2007 through 2015, followed up through 2016, continuously enrolled in Medicaid, and identified from the linked Missouri Cancer Registry and Medicaid claims data set. Data were analyzed from January 2022 to October 2023.

EXPOSURE

Preexisting diabetes.

MAIN OUTCOMES AND MEASURES

Logistic regression was used to estimate odds ratios (ORs) of utilization (yes/no), timely initiation (≤90 days postsurgery), and completion of radiotherapy and chemotherapy, as well as adherence (medication possession ratio ≥80%) and persistence (<90-consecutive day gap) of endocrine therapy in the first year of treatment for women with diabetes compared with women without diabetes. Analyses were adjusted for sociodemographic and tumor factors.

RESULTS

Among 3704 women undergoing definitive surgery, the mean (SD) age was 51.4 (8.6) years, 1038 (28.1%) were non-Hispanic Black, 2598 (70.1%) were non-Hispanic White, 765 (20.7%) had a diabetes history, 2369 (64.0%) received radiotherapy, 2237 (60.4%) had chemotherapy, and 2505 (67.6%) took endocrine therapy. Compared with women without diabetes, women with diabetes were less likely to utilize radiotherapy (OR, 0.67; 95% CI, 0.53-0.86), receive chemotherapy (OR, 0.67; 95% CI, 0.48-0.93), complete chemotherapy (OR, 0.71; 95% CI, 0.50-0.99), and be adherent to endocrine therapy (OR, 0.71; 95% CI, 0.56-0.91). There were no significant associations of diabetes with utilization (OR, 0.95; 95% CI, 0.71-1.28) and persistence (OR, 1.09; 95% CI, 0.88-1.36) of endocrine therapy, timely initiation of radiotherapy (OR, 1.09; 95% CI, 0.86-1.38) and chemotherapy (OR, 1.09; 95% CI, 0.77-1.55), or completion of radiotherapy (OR, 1.25; 95% CI, 0.91-1.71).

CONCLUSIONS AND RELEVANCE

In this cohort study, preexisting diabetes was associated with subpar adjuvant therapies for breast cancer among low-income women. Improving diabetes management during cancer treatment is particularly important for low-income women with breast cancer who may have been disproportionately affected by diabetes and are likely to experience disparities in cancer treatment and outcomes.

Neighborhood Disadvantage, African Genetic Ancestry, Cancer Subtype, and Mortality Among Breast Cancer Survivors

Importance

Racial disparities in breast cancer (BC) survival arise from multilevel causes, which may exert influence at different stages of BC progression. Clarifying the importance of genetic and social factors could help prioritize interventions.

Objective

To jointly examine associations between African genetic ancestry, social environment, and mortality from any cause and BC in Black BC survivors.

Design, Setting, and Participants

This population-based cohort study enrolled self-identified Black women aged 20 to 75 years with histologically confirmed BC from June 2005 to May 2019 and followed them up until death or censoring in September 2021. Participants lived in 10 New Jersey counties. Data were analyzed between December 2022 and April 2023.

Exposures

A neighborhood socioeconomic status (nSES) index composed of census tract measures (education, income, wealth, employment status, and occupation) was linked to residential addresses at diagnosis. Percentage African ancestry was estimated using the ADMIXTURE program.

Main Outcomes and Measures

Sequentially adjusted (age adjusted: age and interview year; fully adjusted: age adjusted with individual SES, lifestyle factors, and comorbidities) logistic regression models were fit to estimate associations with tumor subtypes (estrogen receptor-negative [ER-] vs estrogen receptor-positive [ER+]; triple-negative breast cancer [TNBC] vs luminal A), and Cox models were fit for associations with all-cause mortality (ACM) and breast cancer-specific mortality (BCSM). Models for BCSM were fit using Fine-Gray competing risks models, and robust standard errors were used to account for census tract-level clustering.

Results

Among 1575 participants, median (IQR) African ancestry was 85% (76%-90%), and median (IQR) age was 55 (46-63) years. A 10-percentage point increase in African ancestry was associated with higher odds of ER- vs ER+ (adjusted odds ratio [aOR], 1.08; 95% CI, 0.98-1.18) and TNBC vs luminal (aOR, 1.15; 95% CI, 1.02-1.31) tumors, but not with ACM or BCSM. A 1-IQR increase in nSES was associated with lower ACM (adjusted hazard ratio [aHR], 0.76; 95% CI, 0.63-0.93), and the HR for BCSM was less than 1 but not statistically significant (aHR, 0.81; 95% CI, 0.62-1.04) in age-adjusted models, but associations attenuated following further adjustment for potential mediators (individual SES, lifestyles, comorbidities).

Conclusions and Relevance

In this cohort study of Black female BC survivors, higher African ancestry was associated with aggressive tumor subtypes. Compared with genetic ancestry, mediating pathways related to social environments may be more important for survival in these patients.

Relationships between neighborhood disadvantage and cardiovascular findings at autopsy in subjects with sudden death

BACKGROUND

Neighborhood disadvantage is associated with a higher risk of sudden cardiac death. However, autopsy findings have never been investigated in this context. Here, we sought to explore associations between neighborhood disadvantage and cardiovascular findings at autopsy in cases of sudden death in the State of Maryland.

METHODS

State of Maryland investigation reports from 2,278 subjects within the CVPath Sudden Death Registry were screened for street addresses and 9-digit zip codes. Area deprivation index (ADI), used as metric for neighborhood disadvantage, was available for 1,464 subjects; 650 of whom self-identified as Black and 814 as White. The primary study outcome measurements were causes of death and gross and histopathologic findings of the heart.

RESULTS

Subjects from most disadvantaged neighborhoods (i.e., ADI ≥ 8; n = 607) died at younger age compared with subjects from less disadvantaged neighborhoods (i.e., ADI ≤ 7; n = 857; 46.07 ± 14.10 vs 47.78 ± 13.86 years; P = 0.02) and were more likely Black or women. They were less likely to die from cardiac causes of death (61.8% vs 67.7%; P = 0.02) and had less severe atherosclerotic plaque features, including plaque burden, calcification, intraplaque hemorrhage, and thin-cap fibroatheromas. In addition, subjects from most disadvantaged neighborhoods had lower frequencies of plaque rupture (18.8% vs 25.1%, P = 0.004). However, these associations were omitted after adjustment for traditional risk factors and race.

CONCLUSION

Neighborhood disadvantage did not associate with cause of death or coronary histopathology after adjustment for cardiovascular risk factors and race, implying that social determinants of health other than neighborhood disadvantage play a more prominent role in sudden cardiac death.

Association of Endocrine Therapy Initiation Timeliness With Adherence and Continuation in Low-Income Women With Breast Cancer

IMPORTANCE

Though adjuvant endocrine therapy (AET) has proven efficacy in treating hormone receptor-positive (HR-positive) breast cancer, patient adherence to AET and continuation of treatment as recommended by guidelines remain suboptimal, especially for low-income patients.

OBJECTIVE

To quantify timelines for initiating AET and assess their association with short- and long-term adherence and continuation of AET in low-income women with breast cancer.

DESIGN, SETTING, AND PARTICIPANTS

This population-based retrospective cohort study included women younger than 65 years diagnosed with first primary HR-positive breast cancer between January 1, 2007, and December 31, 2013, followed up for 5 years after the first use of AET through December 2018, and identified from the linked Missouri Cancer Registry and Medicaid claims data set.

EXPOSURES

Time to initiation (TTI) as days from the date of last treatment (surgery, radiotherapy, or chemotherapy) to the first date of AET prescription fill.

MAIN OUTCOMES AND MEASURES

The main outcomes were adherence to AET as medication possession ratio of 80% or greater and continuation of AET as no gap in medication supply for at least 90 days. Odds ratios (ORs) of adherence and continuation over 1 to 5 years were estimated using logistic regression adjusted for demographic, clinical, and neighborhood variables. Analyses were performed between September 1, 2020, and May 31, 2022.

RESULTS

Among 1711 patients, median TTI was 53 (IQR, 26-117) days. A total of 1029 patients (60.1%) were aged 50 to 64 years old, 1270 (74.2%) were non-Hispanic White, and 1133 (66.2%) were unmarried. In the first year after initiation, 1317 (77.0%) were adherent and 1015 (59.3%) continued AET. Over the full 5 years, 376 (22.0%) were adherent and 409 (23.9%) continued AET. Longer TTI was significantly associated with poorer adherence at every year, with an OR of 0.97 (95% CI, 0.95-0.99) for 1-year adherence and an OR of 0.94 (95% CI, 0.90-0.97) for 5-year adherence per 1-month increase in TTI. Longer TTI was also associated with lower odds of short-term, but not long-term, continuation (OR, 0.97 [95% CI, 0.95-0.99] for 1-year continuation and 0.98 [95% CI, 0.96-0.99] for 2-year continuation).

CONCLUSIONS AND RELEVANCE

In this cohort study, longer time to AET initiation was associated with lower odds of short-term and long-term adherence to AET in Medicaid-insured patients with breast cancer. Therefore, early interventions targeting treatment initiation timelines may positively impact adherence throughout the course of treatment and, therefore, outcomes.

Breast cancer incidence and survival in Scotland by socio-economic deprivation and tumour subtype

Background

Women from socio-economically deprived areas are less likely to develop and then to survive breast cancer (BC). Whether associations between deprivation and BC incidence and survival differ by tumour molecular subtypes and mode of detection in Scotland are unknown.

Methods

Data consisted of 62,378 women diagnosed with invasive BC between 2000 and 2016 in Scotland. Incidence rates and time trends were calculated for oestrogen receptor positive (ER+) and negative (ER−) tumours and stratified by the Scottish Index of Multiple Deprivation (SIMD) quintiles and screening status. SIMD is an area-based measure derived across seven domains: income, employment, education, health, access to services, crime and housing. We calculated adjusted hazard ratios (aHR [95% confidence intervals]) for BC death by immunohistochemical surrogates of molecular subtypes for the most versus the least deprived quintile. We adjusted for mode of detection and other confounders.

Results

In Scotland, screen-detected ER+tumour incidence increased over time, particularly in the least deprived quintile [Average Annual Percentage Change (AAPC) = 2.9% with 95% CI from 1.2 to 4.7]. No marked differences were observed for non-screen-detected ER+tumours or ER− tumours by deprivation. BC mortality was higher in the most compared to the least deprived quintile irrespective of ER status (aHR = 1.29 [1.18, 1.41] for ER+ and 1.27 [1.09, 1.47] for ER− tumours). However, deprivation was associated with significantly higher mortality for luminal A and HER2−enriched tumours (aHR = 1.46 [1.13, 1.88] and 2.10 [1.23, 3.59] respectively) but weaker associations for luminal B and TNBC tumours that were not statistically significant.

Conclusions

Deprivation is associated with differential BC incidence trends for screen-detected ER+tumours and with higher mortality for select tumour subtypes. Future efforts should evaluate factors that might be associated with reduced survival in deprived populations and monitor progress stratified by tumour subtypes and mode of detection.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10549-022-06632-1.

Systematic review of neighborhood socioeconomic indices studied across the cancer control continuum

BACKGROUND

There is extensive interest in understanding how neighborhood socioeconomic status (nSES) may affect cancer incidence or survival. However, variability regarding items included and approaches used to form a composite nSES index presents challenges in summarizing overall associations with cancer. Given recent calls for standardized measures of neighborhood sociodemographic effects in cancer disparity research, the objective of this systematic review was to identify and compare existing nSES indices studied across the cancer continuum (incidence, screening, diagnosis, treatment, survival/mortality) and summarize associations by race/ethnicity and cancer site to inform future cancer disparity studies.

METHODS

Using PRISMA guidelines, peer-reviewed articles published between 2010 and 2019 containing keywords related to nSES and cancer were identified in PubMed.

RESULTS

Twenty-four nSES indices were identified from 75 studies. In general, findings indicated a significant association between nSES and cancer outcomes (n = 64/75 studies; 85.33%), with 42/64 (65.63%) adjusting for highly-correlated individual SES factors (e.g., education). However, the direction of association differed by cancer site, race/ethnicity, and nSES index.

CONCLUSIONS

This review highlights several methodologic and conceptual issues surrounding nSES measurement and potential associations with cancer disparities. Recommendations pertaining to the selection of nSES measures are provided, which may help inform disparity-related disease processes and improve the identification of vulnerable populations in need of intervention.

The Effect of Neighborhood Deprivation on Mortality in Newly Diagnosed Diabetes Patients: A Countrywide Population-Based Korean Retrospective Cohort Study, 2002-2013

Background: Neighborhood environmental factors along with individual factors are beginning to make a mark as factors which influence individual health outcomes. The goal of this study is to look at the combined impact of individual and neighborhood socioeconomic status on all-cause mortality in diabetic patients who have just been diagnosed. Methods: The Korean National Health Insurance (2002–2013) was employed in this cohort research, which used a stratified random sample. During the years 2003–2006, a total of 15,882 individuals who were newly diagnosed with diabetes and using oral disease-controlling medication were included in the study. Individual income and neighborhood deprivation index were used to examine the combined effect on all-cause mortality. The frailty model was performed using Cox’s proportional hazard regression. Results: During the study period, 28.3 percent (n = 4493) of the 15,882 eligible individuals died. In a Cox regression analysis after adjusting for all covariates, with advantaged and disadvantaged neighborhoods classified according to individual household income, the adjusted HR for patients living in a disadvantaged area was higher compared to patients living in an advantaged area in patients with middle income, compared to the reference group (a high income within an advantaged neighborhood) (HR, 1.22; 95% CI, 1.09–1.35; HR, 1.13; 95% CI, 1.02–1.25, respectively). The adjusted HR for patients with low income who lived in a disadvantaged location was greater than for patients who lived in an advantaged area (HR, 1.34; 95% CI, 1.18–1.53 vs. HR, 1.28; 95% CI, 1.14–1.49). Conclusions: Individual SES has a greater impact on all-cause mortality among diabetic patients when they live in a low-income neighborhood.

Association between Neighborhood Social Deprivation and Stage at Diagnosis among Breast Cancer Patients in South Carolina

The purpose of this study was to examine the association between neighborhood social deprivation and individual-level characteristics on breast cancer staging in African American and white breast cancer patients. We established a retrospective cohort of patients with breast cancer diagnosed from 1996 to 2015 using the South Carolina Central Cancer Registry. We abstracted sociodemographic and clinical variables from the registry and linked these data to a county-level composite that captured neighborhood social conditions-the social deprivation index (SDI). Data were analyzed using chi-square tests, Student's t-test, and multivariable ordinal regression analysis to evaluate associations. The study sample included 52,803 female patients with breast cancer. Results from the multivariable ordinal regression model demonstrate that higher SDI (OR = 1.06, 95% CI: 1.02-1.10), African American race (OR = 1.35, 95% CI: 1.29-1.41), and being unmarried (OR = 1.17, 95% CI: 1.13-1.22) were associated with a distant stage at diagnosis. Higher tumor grade, younger age, and more recent year of diagnosis were also associated with distant-stage diagnosis. As a proxy for neighborhood context, the SDI can be used by cancer registries and related population-based studies to identify geographic areas that could be prioritized for cancer prevention and control efforts.

Evaluation of Racial/Ethnic Differences in Treatment and Mortality Among Women With Triple-Negative Breast Cancer

Importance

To our knowledge, there is no consensus regarding differences in treatment and mortality between non-Hispanic African American and non-Hispanic White women with triple-negative breast cancer (TNBC). Little is known about whether racial disparities vary by sociodemographic, clinical, and neighborhood factors.

Objective

To examine the differences in clinical treatment and outcomes between African American and White women in a nationally representative cohort of patients with TNBC and further examine the contributions of sociodemographic, clinical, and neighborhood factors to TNBC outcome disparities.

Design, Setting, and Participants

This population-based, retrospective cohort study included 23 123 women who received a diagnosis of nonmetastatic TNBC between January 1, 2010, and December 31, 2015, followed up through December 31, 2016, and identified from the Surveillance, Epidemiology, and End Results data set. The study was conducted from July 2019 to November 2020. The analyses were performed from July 2019 to June 2020.

Exposures

Race and ethnicity, including non-Hispanic African American and non-Hispanic White race.

Main Outcomes and Measures

Using logistic regression analysis and competing risk regression analysis, we estimated odds ratios (ORs) of receipt of treatment and hazard ratios (HRs) of breast cancer mortality in African American patients compared with White patients.

Results

Of 23 213 participants, 5881 (25.3%) were African American women and 17 332 (74.7%) were White women. Compared with White patients, African American patients had lower odds of receiving surgery (OR, 0.69; 95% CI, 0.60-0.79) and chemotherapy (OR, 0.89; 95% CI, 0.81-0.99) after adjustment for sociodemographic, clinicopathologic, and county-level factors. During a 43-month follow-up, 3276 patients (14.2%) died of breast cancer. The HR of breast cancer mortality was 1.28 (95% CI, 1.18-1.38) for African American individuals after adjustment for sociodemographic and county-level factors. Further adjustment for clinicopathological and treatment factors reduced the HR to 1.16 (95% CI, 1.06-1.25). This association was observed in patients living in socioeconomically less deprived counties (HR, 1.26; 95% CI, 1.14-1.39), urban patients (HR, 1.21; 95% CI, 1.11-1.32), patients having stage II (HR, 1.19; 95% CI, 1.02-1.39) or III (HR, 1.15; 95% CI, 1.01-1.31) tumors that were treated with chemotherapy, and patients younger than 65 years (HR, 1.24; 95% CI, 1.12-1.37).

Conclusions and Relevance

In this retrospective cohort study, African American women with nonmetastatic TNBC had a significantly higher risk of breast cancer mortality compared with their White counterparts, which was partially explained by their disparities in receipt of surgery and chemotherapy.

Neighborhood archetypes and breast cancer survival in California

PURPOSE

Previous studies on neighborhoods and breast cancer survival examined neighborhood variables as unidimensional measures (e.g. walkability or deprivation) individually and thus cannot inform how the multitude of highly correlated neighborhood domains interact to impact breast cancer survival. Neighborhood archetypes were developed that consider interactions among a broad range of neighborhood social and built environment attributes and examine their associations with breast cancer survival.

METHODS

Archetypes were measured using latent class analysis (LCA) fit to California census tract-level data. Thirty-nine social and built environment attributes relevant to eight neighborhood domains (socioeconomic status (SES), urbanicity, demographics, housing, land use, commuting and traffic, residential mobility, and food environment) were included.  The archetypes were linked to cancer registry data on breast cancer cases (diagnosed 1996-2005 with follow-up through Dec 31, 2017) to evaluate their associations with overall and breast cancer-specific survival using Cox proportional hazards models. Analyses were stratified by race/ethnicity.

RESULTS

California neighborhoods were best described by nine archetypal patterns that were differentially associated with overall and breast cancer-specific survival. The lowest risk of overall death was observed in the upper middle class suburb (reference) and high status neighborhoods, while the highest was observed among inner city residents with a 39% greater risk of death (95% CI = 1.35 to 1.44). Results were similar for breast cancer-specific survival. Stratified analyses indicated that differences in survival by neighborhood archetypes varied according to individuals' race/ethnicity.

CONCLUSIONS

By describing neighborhood archetypes that differentiate survival following breast cancer diagnosis, the study provides direction for policy and clinical practice addressing contextually-rooted social determinants of health including SES, unhealthy food environments, and greenspace.

Relationship between insurance status and outcomes for patients with breast cancer in Missouri

BACKGROUND

The cancer stage at diagnosis, treatment delays, and breast cancer mortality vary with insurance status.

METHODS

Using the Missouri Cancer Registry, this analysis included 31,485 women diagnosed with invasive breast cancer from January 1, 2007, to December 31, 2015. Odds ratios (ORs) of a late-stage (stage III or IV) diagnosis and a treatment delay (>60 days after the diagnosis) were calculated with logistic regression. The hazard ratio (HR) of breast cancer mortality was calculated with Cox proportional hazards regression. Mediation analysis was used to quantify the individual contributions of each covariate to mortality.

RESULTS

The OR of a late-stage diagnosis was higher for patients with Medicaid (OR, 1.72; 95% confidence interval [CI], 1.56-1.91) or no insurance (OR, 2.30; 95% CI, 1.91-2.78) in comparison with privately insured patients. Medicare (OR, 1.21; 95% CI, 1.10-1.37), Medicaid (OR, 1.60; 95% CI, 1.37-1.85), and uninsured patients (OR, 1.58; 95% CI, 1.18-2.12) had higher odds of a treatment delay. The HR of breast cancer-specific mortality was significantly increased in the groups with public insurance or no insurance and decreased after sequential adjustments for sociodemographic factors (HR, 2.39; 95% CI, 1.96-2.91), tumor characteristics (HR, 1.28; 95% CI, 1.05-1.56), and treatment (HR, 1.23; 95% CI, 1.01-1.50). Late-stage diagnoses accounted for 72.5% of breast cancer mortality in the uninsured.

CONCLUSIONS

Compared with the privately insured, women with public or no insurance had a higher risk for advanced breast cancer, a >60-day treatment delay, and death from breast cancer. Particularly for the uninsured, Medicaid expansion and increased funding for education and screening programs could decrease breast cancer disparities.

Racial and ethnic disparities in 21-gene recurrence scores, chemotherapy, and survival among women with hormone receptor-positive, node-negative breast cancer

PURPOSE

Cutoffs of the 21-gene recurrence score (RS), a commonly used genomic assay for hormone receptor-positive breast cancer, have been updated. Little is known about racial/ethnic differences in RS results, RS-guided chemotherapy use, and outcomes on updated cutoff (RS ≥ 31 defined as high-risk) in the real-world setting.

METHODS

A total of 81,937 women [75.0% whites, 7.7% blacks, 8.3% Asian American/Pacific Islanders (AAPIs), and 9.0% Hispanics] diagnosed with hormone receptor-positive breast cancer between 2004 and 2015, who received the 21-gene assay, were identified from the Surveillance, Epidemiology, and End Results. Logistic regressions estimated the race-associated odds ratios (ORs) of RS and chemotherapy use. Cox regressions estimated the race-associated hazard ratios (HRs) of breast cancer-specific and all-cause mortality.

RESULTS

Compared with white women, black women were more likely to have RS-defined high-risk tumors (adjusted OR [aOR] 1.29; 95% CI 1.16-1.42). In high RS, blacks had lower odds of chemotherapy use (aOR 0.76; 95% CI 0.62-0.94) than whites, particularly among women ≥ 65 years (aOR 0.51; 95% CI 0.35-0.76), while AAPI and Hispanic women had no variation in chemotherapy use compared with whites in high RS. Black women had a higher risk of breast cancer-specific mortality (HR 1.37; 95% CI 1.12-1.67) and all-cause mortality compared with white women after adjusting for demographic and pathological factors, county-level socioeconomic deprivation, treatments and RS; AAPIs had lower mortality and Hispanics had similar mortality.

CONCLUSIONS

Black women were more likely to have a high-risk RS tumor and less likely to receive chemotherapy in the group of high RS, especially those ≥ 65 years. Further studies are needed to identify barriers to chemotherapy in black patients with high RS scores.

A Novel Prediction Tool Based on Large Cohorts to Determine the Cancer-Specific Survival Probability of Patients With Locally Advanced Pancreatic Cancer After Irreversible Electroporation Treatment

Irreversible electroporation (IRE) is a novel method which was especially suitable for the treatment of locally advanced pancreatic cancer (LAPC). The purpose of this study was to evaluate probabilities of overall survival (OS) and cancer-specific survival (CSS) in patients with LAPC after IRE treatment and to construct nomograms to predict survival for these patients. Data of patients were retrospectively collected from the Surveillance, Epidemiology, and End Results (SEER) database and medical records of Sun Yat-sen University Cancer Center (SYSUCC). A total of 312 LAPC patients after IRE treatment were included into this study. The 3-year cumulative incidence of cancer-specific mortality for patients with LAPC after IRE treatment was 74.3%. Nomograms for predicting probabilities of OS, CSS, and non-cancer-specific survival (NCSS) were built and calibrated with the concordance index (C-index) and the area under receiver operating characteristic (ROC) curve (AUC). The established nomograms were well-calibrated, with C-indexes of 0.782 for OS prediction, 0.729 for CSS prediction, and 0.730 for NCSS prediction. Compared with the TNM stage system, the established nomograms displayed higher values of AUC and showed better discriminatory power for predicting OS, CSS, and NCSS. These nomograms were well-calibrated and could serve to guide management of LAPC patients after IRE treatment.

Influence of built environment on quality of life changes in African-American patients with non-metastatic breast cancer

Research links the built environment to health outcomes, but little is known about how this affects quality of life (QOL) of African American breast cancer patients, especially those residing in disadvantaged neighborhoods. Using latent trajectory models, we examined whether the built environment using Google Street View was associated with changes in QOL over a 2-year follow-up in 228 newly diagnosed African American breast cancer patients. We measured QOL using the RAND 36-Item Health Survey subscales. After adjusting for covariates, improvement in emotional well-being and pain over time was greater for women living on streets with low-quality (vs. high-quality) sidewalks.

Factors associated with HPV testing in oropharyngeal cancer in the National Cancer Data Base from 2013 to 2015

OBJECTIVES

Given the recent treatment deintensification clinical trials for Human Papillomavirus (HPV)-associated oropharyngeal cancer, College of American Pathologists recommendation for HPV-testing of all oropharyngeal cancers and treatment disparities in head and neck cancer, determining factors related to HPV testing are exceedingly pertinent.

MATERIALS AND METHODS

We used the National Cancer Data Base, accounting for 70% of new cancer diagnoses. To reduce the heterogeneity due to the recent recommendation of HPV-testing, we used squamous cell oropharyngeal cancer patients from 2013 to 2015. We only used patients that have either reported HPV testing and non-testing (30.5% of the sample). We used a chi-square test to compare the factors among tested and untested patients and calculated the prevalence ratio for not tested to those tested with Poisson regression. As a sensitivity analysis, we used a fully Conditional Specification implemented by the MICE algorithm to impute missing variables.

RESULTS

Of the 24,241 oropharyngeal cancer patients with HPV testing data, 12% were not been tested for HPV. Across the study period, integrated network and low-volume hospitals had the lowest proportion of HPV testing from 2013 to 2015. In a multivariable analysis, compared to patients with private insurance, Medicaid (PR: 1.82; 95% Confidence Interval (CI): 1.63-2.02) and uninsured (PR: 1.75; 95% CI: 1.52-2.01) patients were more likely not to be tested for HPV. We saw similar results in the imputed dataset, in which 12.5% of patients were not tested.

CONCLUSIONS

This heterogeneity in testing is significant, given potential de-intensification of treatment for HPV-positive cancer. Future research should examine interventions in non-academic low-volume to ensure equitable treatment for all.

Neighborhood Characteristics and Colorectal Cancer Survivors' Quality of Care

Purpose: Quality cancer care entails receipt of a Survivorship Care Plan (SCP). The purpose of this study was to determine differences in SCP delivery by patient-level and neighborhood characteristics.

Methods: We obtained California cancer registry data on individuals who were diagnosed with stage I, II, or III colorectal cancer (CRC) between 2012 and 2015 and resided in predetermined geographic areas. We then mailed them a questionnaire, which queried about receipt of a SCP and its content. SCP was defined by content, as summary of cancer treatment, cancer surveillance recommendations, and/or an individualized preventive care. Using logistic regression modeling, each measure of SCP, as well as the summary measure (none vs. any), was evaluated by person-level characteristics. Subsequently, neighborhood-level characteristics were added to the model to explore their additional value.

Results: Overall 80% of CRC survivors received a SCP. Receipt of SCPs was associated with person-level characteristics, while neighborhood characteristics did not make an additional contribution. Young, male employed survivors and those with more recent diagnoses or later cancer stages had greater odds of receiving a SCP.

Conclusion: When providing SCPs, health care providers prioritize patient groups who they may perceive as vulnerable or likely to benefit from SCPs.

Measuring Subcounty Differences in Population Health Using Hospital and Census-Derived Data Sets: The Missouri ZIP Health Rankings Project

CONTEXT

Measures of population health at the subcounty level are needed to identify areas for focused interventions and to support local health improvement activities.

OBJECTIVE

To extend the County Health Rankings population health measurement model to the ZIP code level using widely available hospital and census-derived data sources.

DESIGN

Retrospective administrative data study.

SETTING

Missouri.

POPULATION

Missouri FY 2012-2014 hospital inpatient, outpatient, and emergency department discharge encounters (N = 36 176 377) and 2015 Nielsen data.

MAIN OUTCOME MEASURES

ZIP code-level health factors and health outcomes indices.

RESULTS

Statistically significant measures of association were observed between the ZIP code-level population health indices and published County Health Rankings indices. Variation within counties was observed in both urban and rural areas. Substantial variation of the derived measures was observed at the ZIP code level with 20 (17.4%) Missouri counties having ZIP codes in both the top and bottom quintiles of health factors and health outcomes. Thirty of the 46 (65.2%) counties in the top 2 county quintiles had ZIP codes in the bottom 2 quintiles.

CONCLUSIONS

This proof-of-concept analysis suggests that readily available hospital and census-derived data can be used to create measures of population health at the subcounty level. These widely available data sources could be used to identify areas of potential need within counties, engage community stakeholders, and target interventions.

Competing risk analyses of overall survival and cancer-specific survival in patients with combined hepatocellular cholangiocarcinoma after surgery

Background

Our objective was to identify risk factors affecting overall survival (OS) and cancer-specific survival (CSS) and build nomograms to predict survival based on a large population-based cohort.

Methods

Two hundred and thirty patients diagnosed with CHCC between 2004 and 2015 were retrospectively extracted from the Surveillance, Epidemiology, and End Results (SEER) database as a training cohort. In addition, Ninety-nine patients diagnosed with CHCC between 2000 and 2017 were retrospectively extracted from Sun Yat-Sen University Cancer Center (SYSUCC) as an external validation. Nomograms for predicting probability of OS and CSS were established. Performance of the nomograms was measured by concordance index (C-index) and the area under receiver operating characteristic (ROC) curve (AUC).

Results

In training cohort, the 1-, 2 and 3-year OS were 67.7, 46.8 and 37.9%, and the 1-, 2 and 3-year CSS were 73.1, 52.0 and 43.0%, respectively. The established nomograms were well calibrated in both training and validation cohort, with concordance indexes (C-index) of 0.652 and 0.659, respectively for OS prediction; 0.706 and 0.763, respectively for CSS prediction. Nomograms also displayed better discriminatory compared with 8th edition tumor-node-metastasis (TNM) stage system for predicting OS and CSS.

Conclusion

We constructed nomograms to predict OS and CSS based on a relatively large cohort. The established nomograms were well validated and could serve to improve predictions of survival risks and guide management of patients with CHCC after surgery.

Impacts of Neighborhood Characteristics on Treatment and Outcomes in Women with Ductal Carcinoma In Situ of the Breast

Background: This study examines associations of neighborhood characteristics with treatment and outcomes of ductal carcinoma in situ (DCIS) of the breast.Methods: From the Missouri Cancer Registry, we identified 9,195 women with DCIS diagnosed between 1996 and 2011. A composite index using U.S. Census data and American Community Survey data was developed to assess census tract-level socioeconomic deprivation, and rural-urban commuting area codes were used to define rural census tracts. ORs and 95% confidence intervals (CIs) of the treatment were estimated using logistic regression. Hazard ratios (HRs) of DCIS outcomes were estimated using Cox proportional hazards regression.Results: Women in the most socioeconomically deprived census tracts were more likely than those in the least deprived to have mastectomy (OR = 1.44; 95% CI, 1.25-1.66; P _trend < 0.0001), no surgery (OR = 1.54; 95% CI, 1.02-2.30; P _trend = 0.04), no radiotherapy post-breast conserving surgery (OR = 1.90; 95% CI, 1.56-2.31; P _trend<0.0001), delayed radiotherapy (OR = 1.26; 95% CI, 1.01-1.57; P _trend = 0.02), and ipsilateral breast tumors (HR = 1.59; 95% CI, 1.07-2.38; P _trend = 0.03). There was no significant difference in risk of contralateral breast tumors. Compared with urban women, rural women had significantly higher odds of underutilization of radiotherapy (OR = 1.29; 95% CI, 1.08-1.53). Rural locations were not associated with risk of ipsilateral or contralateral breast tumors.Conclusions: Neighborhood socioeconomic deprivation was associated with higher risks of suboptimal treatment and ipsilateral breast tumors. While DCIS treatment significantly varied by rural/urban locations, we did not observe any statistically significant rural-urban differences in risks of second breast tumors.Impact: Neighborhood attributes may affect treatment and outcomes of patients with DCIS. Cancer Epidemiol Biomarkers Prev; 27(11); 1298-306. ©2018 AACR.

Literature review of data-based models for identification of factors associated with racial disparities in breast cancer mortality

In the United States, early detection methods have contributed to the reduction of overall breast cancer mortality but this pattern has not been observed uniformly across all racial groups. A vast body of research literature shows a set of health care, socio-economic, biological, physical, and behavioural factors influencing the mortality disparity. In this paper, we review the modelling frameworks, statistical tests, and databases used in understanding influential factors, and we discuss the factors documented in the modelling literature. Our findings suggest that disparities research relies on conventional modelling and statistical tools for quantitative analysis, and there exist opportunities to implement data-based modelling frameworks for (1) exploring mechanisms triggering disparities, (2) increasing the collection of behavioural data, and (3) monitoring factors associated with the mortality disparity across time.

Urban Neighborhood and Residential Factors Associated with Breast Cancer in African American Women: a Systematic Review

Residential characteristics in urban neighborhoods impact health and might be important factors contributing to health disparities, especially in the African American population. The purpose of this systematic review is to understand the relationship between urban neighborhood and residential factors and breast cancer incidence and prognosis in African American women. Using PubMed and Web of Science, the existing literature was reviewed. Observational, cross-sectional, cohort, and prospective studies until February 2017 were examined. Studies including populations of African American women, setting in "urban" areas, and a measure of a neighborhood or residential factor were reviewed. Four parameters related to neighborhood or residential factors were extracted including: neighborhood socioeconomic status (nSES), residential segregation, spatial access to mammography, and residential pollution. Our analysis showed that African American women living in low nSES have greater odds of late stage diagnosis and mortality. Furthermore, African American women living in segregated areas (higher percentage of Blacks) have higher odds of late stage diagnosis and mortality compared to White and Hispanic women living in less segregated areas (lower percentage of Blacks). Late stage diagnosis was also shown to be significantly higher in areas with poor mammography access and areas with higher Black residential segregation. Lastly, residential pollution did not affect breast cancer risk in African American women. Overall, this systematic review provides a qualitative synthesis of major neighborhood and residential factors on breast cancer outcomes in African American women.

Contextual effects and cancer outcomes in the United States: a systematic review of characteristics in multilevel analyses

PURPOSE

There is increasing call for the utilization of multilevel modeling to explore the relationship between place-based contextual effects and cancer outcomes in the United States. To gain a better understanding of how contextual factors are being considered, we performed a systematic review.

METHODS

We reviewed studies published between January 1, 2002 and December 31, 2016 and assessed the following attributes: (1) contextual considerations such as geographic scale and contextual factors used; (2) methods used to quantify contextual factors; and (3) cancer type and outcomes. We searched PubMed, Scopus, and Web of Science and initially identified 1060 studies. One hundred twenty-two studies remained after exclusions.

RESULTS

Most studies utilized a two-level structure; census tracts were the most commonly used geographic scale. Socioeconomic factors, health care access, racial/ethnic factors, and rural-urban status were the most common contextual factors addressed in multilevel models. Breast and colorectal cancers were the most common cancer types, and screening and staging were the most common outcomes assessed in these studies.

CONCLUSIONS

Opportunities for future research include deriving contextual factors using more rigorous approaches, considering cross-classified structures and cross-level interactions, and using multilevel modeling to explore understudied cancers and outcomes.

Geographic variability in gestational weight gain: a multilevel population-based study of women having term births in Florida (2005-2012)

PURPOSE

We examined the extent of geographic variability in gestational weight gain (GWG), identified areas where women have suboptimal GWG, and evaluated whether individual- and area-level factors account for such variability.

METHODS

We conducted a population-based cohort study including 1,385,574 women delivering term, singleton, and live births in Florida. We used a Bayesian, structured additive regression with a spatial function to analyze data from Florida's birth certificates (2005-2012) and ZIP code tabulation areas (ZCTAs; 2010 Census).

RESULTS

The prevalence of insufficient (7.7%-42.9%) and excessive (17.1%-82.4%) GWG varied widely within Florida. Geographic variability was not explained by risk factors under study. Clusters in Orlando, Tampa, and Miami exhibited increased likelihood of insufficient GWG, whereas clusters in the Northwest of Florida exhibited increased likelihood of excessive GWG.

CONCLUSIONS

We identified areas in Florida with high likelihood of suboptimal GWG that policy-makers should prioritize in the implementation of programs for optimizing GWG.

The intersection of aggregate-level lead exposure and crime

CONTEXT

Childhood lead exposure has been associated with criminal behavior later in life. The current study aimed to analyze the association between elevated blood lead levels (n=59,645) and crime occurrence (n=90,433) across census tracts within St. Louis, Missouri.

DESIGN

Longitudinal ecological study.

SETTING

Saint Louis, Missouri.

EXPOSURE MEASURE

Blood lead levels.

MAIN OUTCOME MEASURE

Violent, Non-violent, and total crime at the census tract level.

RESULTS

Spatial statistical models were used to account for the spatial autocorrelation of the data. Greater lead exposure at the census-tract level was associated with increased violent, non-violent, and total crime. In addition, we examined whether non-additive effects existed in the data by testing for an interaction between lead exposure and concentrated disadvantage. Some evidence of a negative interaction emerged, however, it failed to reach traditional levels of statistical significance (supplementary models, however, revealed a similar negative interaction that was significant).

CONCLUSIONS

More precise measurements of lead exposure in the aggregate, produced additional evidence that lead is a potent predictor of criminal outcomes.

Geographic Variation in Maternal Smoking during Pregnancy in the Missouri Adolescent Female Twin Study (MOAFTS)

Objective

Despite well-known adverse health effects of maternal smoking during pregnancy (MSP), it is still unclear if MSP varies geographically and if neighborhood socioeconomic deprivation (SED) plays an important role in MSP. This study aims to investigate small-area geographic variation in MSP and examine the association of SED with MSP.

Methods

The Missouri Adolescent Female Twin Study (MOAFTS) is a cohort study of female like-sex twins born in Missouri to Missouri-resident parents during 1975–1985. Biological mothers completed a baseline interview in 1995–1998 and reported MSP with the twins. Residential address of the mother at birth was geocoded. We developed a census tract-level SED index using a common factor approach based on 21 area-level socioeconomic variables from the 1980 Census data. Multilevel logistic regressions estimated geographic heterogeneity (random effect) in MSP and the odds ratios (ORs, fixed effects) of neighborhood SED associated with MSP.

Results

Of 1658 MOAFTS mothers, 35.2% reported any MSP and 21.9% reported MSP beyond the first trimester. Neighborhood SED was associated with any MSP (the highest vs. the lowest quartile: OR = 1.90, 95% confidence interval [CI] = 1.40–2.57, P_trend<0.001) and MSP beyond the first trimester (OR = 1.98, 95% CI = 1.38–2.85, P_trend = 0.002) in unadjusted analyses. After adjusting for individual covariates (demographics, socioeconomic conditions, alcohol use, and parents’ cohabitation), neighborhood SED was not associated with MSP, but geographic variation still persisted in MSP (variance = 0.41, P = 0.003) and in MSP beyond the first trimester (variance = 0.82, P<0.001).

Conclusions

Neighborhood SED was associated with MSP in unadjusted analyses but this association could be explained by individual socioeconomic conditions. Nonetheless, significant geographic variation in MSP persisted and was not accounted for by differences in neighborhood SED. To develop effective interventions to reduce MSP, further studies are necessary to explore underlying reasons for its geographic variation.

Influence of neighborhood-level factors on social support in early-stage breast cancer patients and controls

RATIONALE

Low social support has been linked to negative health outcomes in breast cancer patients.

OBJECTIVE

We examined associations between perceived social support, neighborhood socioeconomic deprivation, and neighborhood-level social support in early-stage breast cancer patients and controls.

METHODS

This two-year longitudinal study in the United States included information collected from telephone interviews and clinical records of 541 early-stage patients and 542 controls recruited from 2003 to 2007. Social support was assessed using the Medical Outcomes Study Social Support Survey (MOS-SS). Residential addresses were geocoded and used to develop measures including neighborhood social support (based on MOS-SS scores from nearby controls) and neighborhood socioeconomic deprivation (a composite index of census tract characteristics). Latent trajectory models were used to determine effects of neighborhood conditions on the stable (intercept) and changing (slope) aspects of social support.

RESULTS

In a model with only neighborhood variables, greater socioeconomic deprivation was associated with patients' lower stable social support (standardized estimate = -0.12, p = 0.027); neighborhood-level social support was associated with social support change (standardized estimate = 0.17, p = 0.046). After adding individual-level covariates, there were no direct neighborhood effects on social support. In patients, neighborhood socioeconomic deprivation was associated with support indirectly through marriage, insurance status, negative affect, and general health. In controls, neighborhood socioeconomic deprivation was associated with support indirectly through marriage (p < 0.05).

CONCLUSION

Indirect effects of neighborhood socioeconomic deprivation on social support differed in patients and controls. Psychosocial and neighborhood interventions may help patients with low social support, particularly patients without partnered relationships in deprived areas.