Clinical factors associated with patterns of endocrine therapy adherence in premenopausal breast cancer patients.

INTRODUCTION: Patients with hormone receptor positive breast cancer are recommended at least five years of adjuvant endocrine therapy, but adherence to this treatment is often suboptimal. We investigated longitudinal trends in adjuvant endocrine therapy (AET) adherence among premenopausal breast cancer patients and identified clinical characteristics, including baseline comorbidities and non-cancer chronic medication use, associated with AET adherence. METHODS: We included stage I-III premenopausal breast cancer patients diagnosed during 2002-2011 and registered in the Danish Breast Cancer Group clinical database who initiated AET. We used group-based trajectory modeling to describe AET adherence patterns. We also linked patients to Danish population-based registries and fit multinomial logistic models to compute odds ratios (ORs) and 95% confidence intervals (95% CIs) associating clinical characteristics with AET adherence patterns. RESULTS: We identified three adherence patterns among 4,353 women-high adherers (57%), slow decliners (36%), and rapid decliners (6.9%). Women with stage I disease (vs. stage II; OR: 1.9, 95% CI 1.5, 2.5), without chemotherapy (vs. chemotherapy; OR: 4.3, 95% CI 3.0, 6.1), with prevalent comorbid disease (Charlson Comorbidity Index score ≥ 1 vs. 0; OR: 1.6, 95% CI 1.1, 2.3), and with a history of chronic non-cancer medication use (vs. none; OR: 1.3, 95% CI 1.0, 1.8) were more likely to be rapid decliners compared with high adherers. CONCLUSIONS: Women with stage I cancer, no chemotherapy, higher comorbidity burden, and history of chronic non-cancer medication use were less likely to adhere to AET. Taking steps to promote adherence in these groups of women may reduce their risk of recurrence.

Non-differential misclassification of outcome under (near)-perfect specificity: a simulation study.

Mismeasurement of a dichotomous outcome yields an unbiased risk ratio estimate when there are no false positive cases (perfect specificity) and when sensitivity is non-differential with respect to exposure status. In studies where these conditions are expected, quantitative bias analysis may be considered unnecessary. We conducted a simulation study to explore the robustness of this special case to small departures from perfect specificity and stochastic departures from non-differential sensitivity. We observed substantial bias of the risk ratio with specificity values as high at 99.8%. The magnitude of bias increased directly with the true underlying risk ratio and was markedly stronger at lower baseline risk. Stochastic departure from non-differential sensitivity also resulted in substantial bias in most simulated scenarios; downward bias prevailed when sensitivity was higher among unexposed compared with exposed, and upward bias prevailed when sensitivity was higher among exposed compared with unexposed. Our results show that seemingly innocuous departures from perfect specificity (e.g., 0.2%) and from non-differential sensitivity can yield substantial bias of the risk ratio under outcome misclassification. We present a web tool permitting easy exploration of this bias mechanism under user-specifiable study scenarios.

Racial disparities in initiation of chemotherapy among breast cancer patients with discretionary treatment indication in the state of Georgia.

PURPOSE: The majority of breast cancer patients are diagnosed with early-stage estrogen receptor (ER) positive disease. Despite effective treatments for these cancers, Black women have higher mortality than White women. We investigated demographic and clinical factors associated with receipt of chemotherapy among those with a discretionary indication who are at risk for overtreatment. METHODS: Using Georgia Cancer Registry data, we identified females diagnosed with ER positive breast cancer who had a discretionary indication for chemotherapy (2010-2017). We used logistic regression to estimate odds ratios (ORs) and 95% confidence intervals (CIs) associating patient demographic and clinical characteristics with chemotherapy initiation overall, and comparing non-Hispanic Black (NHB) with non-Hispanic White (NHW) women within strata of patient factors. RESULTS: We identified 11,993 ER positive breast cancer patients with a discretionary indication for chemotherapy. NHB patients were more likely to initiate chemotherapy compared with NHW women (OR = 1.41, 95% CI: 1.28, 1.56). Race differences in chemotherapy initiation were pronounced among those who did not receive Oncotype DX testing (OR = 1.47, 95% CI: 1.31, 1.65) and among those residing in high socioeconomic status neighborhoods (OR = 2.48, 95% CI: 1.70, 3.61). However, we observed equitable chemotherapy receipt among patients who received Oncotype DX testing (OR = 0.90, 95% CI: 0.71, 1.14), were diagnosed with grade 1 disease (OR = 1.00, 95% CI: 0.74, 1.37), and those resided in rural areas (OR = 1.01, 95% CI: 0.76, 1.36). CONCLUSION: We observed racial disparities in the initiation of chemotherapy overall and by sociodemographic and clinical factors, and more equitable outcomes when clinical guidelines were followed.

The Population-level Effect of Adjuvant Therapies on Breast Cancer Recurrence: Application of the Trend-in-Trend Design.

PURPOSE: Breast cancer has an average 10-year relative survival reaching 84%. This favorable survival is due, in part, to the introduction of biomarker-guided therapies. We estimated the population-level effect of the introduction of two adjuvant therapies-tamoxifen and trastuzumab-on recurrence using the trend-in-trend pharmacoepidemiologic study design. METHODS: We ascertained data on women diagnosed with nonmetastatic breast cancer who were registered in the Danish Breast Cancer Group clinical database. We used the trend-in-trend design to estimate the population-level effect of the introduction of (1) tamoxifen for postmenopausal women with estrogen receptor (ER)-positive breast cancer in 1982, (2) tamoxifen for premenopausal women diagnosed with ER-positive breast cancer in 1999, and (3) trastuzumab for women <60 years diagnosed with human epidermal growth factor receptor 2-positive breast cancer in 2007. RESULTS: For the population-level effect of the introduction of tamoxifen among premenopausal women diagnosed with ER-positive breast cancer in 1999, the risk of recurrence decreased by nearly one-half (OR = 0.52), consistent with evidence from clinical trials; however, the estimate was imprecise (95% confidence interval [CI] = 0.25, 1.85). We observed an imprecise association between tamoxifen use and recurrence from the time it was introduced in 1982 (OR = 1.24 95% CI = 0.46, 5.11), inconsistent with prior knowledge from clinical trials. For the introduction of trastuzumab in 2007, the estimate was also consistent with trial evidence, though imprecise (OR = 0.51; 95% CI = 0.21, 22.4). CONCLUSIONS: We demonstrated how novel pharmacoepidemiologic analytic designs can be used to evaluate the routine clinical care and effectiveness of therapeutic advancements in a population-based setting while considering some limitations of the approach.

Endometriosis Typology and Ovarian Cancer Risk.

Importance: Endometriosis has been associated with an increased risk of ovarian cancer; however, the associations between endometriosis subtypes and ovarian cancer histotypes have not been well-described. Objective: To evaluate the associations of endometriosis subtypes with incidence of ovarian cancer, both overall and by histotype. Design, Setting, and Participants: Population-based cohort study using data from the Utah Population Database. The cohort was assembled by matching 78 893 women with endometriosis in a 1:5 ratio to women without endometriosis. Exposures: Endometriosis cases were identified via electronic health records and categorized as superficial endometriosis, ovarian endometriomas, deep infiltrating endometriosis, or other. Main Outcomes and Measures: Estimated adjusted hazard ratios (aHRs), adjusted risk differences (aRDs) per 10 000 women, and 95% CIs for overall ovarian cancer, type I ovarian cancer, and type II ovarian cancer comparing women with each type of endometriosis with women without endometriosis. Models accounted for sociodemographic factors, reproductive history, and past gynecologic operations. Results: In this Utah-based cohort, the mean (SD) age at first endometriosis diagnosis was 36 (10) years. There were 597 women with ovarian cancer. Ovarian cancer risk was higher among women with endometriosis compared with women without endometriosis (aHR, 4.20 [95% CI, 3.59-4.91]; aRD, 9.90 [95% CI, 7.22-12.57]), and risk of type I ovarian cancer was especially high (aHR, 7.48 [95% CI, 5.80-9.65]; aRD, 7.53 [95% CI, 5.46-9.61]). Ovarian cancer risk was highest in women with deep infiltrating endometriosis and/or ovarian endometriomas for all ovarian cancers (aHR, 9.66 [95% CI, 7.77-12.00]; aRD, 26.71 [95% CI, 20.01-33.41]), type I ovarian cancer (aHR, 18.96 [95% CI, 13.78-26.08]; aRD, 19.57 [95% CI, 13.80-25.35]), and type II ovarian cancer (aHR, 3.72 [95% CI, 2.31-5.98]; aRD, 2.42 [95% CI, -0.01 to 4.85]). Conclusions and Relevance: Ovarian cancer risk was markedly increased among women with ovarian endometriomas and/or deep infiltrating endometriosis. This population may benefit from counseling regarding ovarian cancer risk and prevention and could be an important population for targeted screening and prevention studies.

Association of inflammation-related exposures and ovarian cancer survival in a multi-site cohort study of Black women.

BACKGROUND: An association was observed between an inflammation-related risk score (IRRS) and worse overall survival (OS) among a cohort of mostly White women with invasive epithelial ovarian cancer (EOC). Herein, we evaluated the association between the IRRS and OS among Black women with EOC, a population with higher frequencies of pro-inflammatory exposures and worse survival. METHODS: The analysis included 592 Black women diagnosed with EOC from the African American Cancer Epidemiology Study (AACES). Cox proportional hazards models were used to compute hazard ratios (HRs) and 95% confidence intervals (CIs) for the association of the IRRS and OS, adjusting for relevant covariates. Additional inflammation-related exposures, including the energy-adjusted Dietary Inflammatory Index (E-DIITM), were evaluated. RESULTS: A dose-response trend was observed showing higher IRRS was associated with worse OS (per quartile HR: 1.11, 95% CI: 1.01-1.22). Adding the E-DII to the model attenuated the association of IRRS with OS, and increasing E-DII, indicating a more pro-inflammatory diet, was associated with shorter OS (per quartile HR: 1.12, 95% CI: 1.02-1.24). Scoring high on both indices was associated with shorter OS (HR: 1.54, 95% CI: 1.16-2.06). CONCLUSION: Higher levels of inflammation-related exposures were associated with decreased EOC OS among Black women.

Role of neighborhood context in ovarian cancer survival disparities: current research and future directions.

Ovarian cancer is the fifth leading cause of cancer-associated mortality among US women with survival disparities seen across race, ethnicity, and socioeconomic status, even after accounting for histology, stage, treatment, and other clinical factors. Neighborhood context can play an important role in ovarian cancer survival, and, to the extent to which minority racial and ethnic groups and populations of lower socioeconomic status are more likely to be segregated into neighborhoods with lower quality social, built, and physical environment, these contextual factors may be a critical component of ovarian cancer survival disparities. Understanding factors associated with ovarian cancer outcome disparities will allow clinicians to identify patients at risk for worse outcomes and point to measures, such as social support programs or transportation aid, that can help to ameliorate such disparities. However, research on the impact of neighborhood contextual factors in ovarian cancer survival and in disparities in ovarian cancer survival is limited. This commentary focuses on the following neighborhood contextual domains: structural and institutional context, social context, physical context represented by environmental exposures, built environment, rurality, and healthcare access. The research conducted to date is presented and clinical implications and recommendations for future interventions and studies to address disparities in ovarian cancer outcomes are proposed.

Sampling Validation Data to Achieve a Planned Precision of the Bias-Adjusted Estimate of Effect.

Data collected from a validation substudy permit calculation of a bias-adjusted estimate of effect that is expected to equal the estimate that would have been observed had the gold standard measurement been available for the entire study population. In this paper, we develop and apply a framework for adaptive validation to determine when sufficient validation data have been collected to yield a bias-adjusted effect estimate with a prespecified level of precision. Prespecified levels of precision are decided a priori by the investigator, based on the precision of the conventional estimate and allowing for wider confidence intervals that would still be substantively meaningful. We further present an applied example of the use of this method to address exposure misclassification in a study of transmasculine/transfeminine youth and self-harm. Our method provides a novel approach to effective and efficient estimation of classification parameters as validation data accrue, with emphasis on the precision of the bias-adjusted estimate. This method can be applied within the context of any parent epidemiologic study design in which validation data will be collected and modified to meet alternative criteria given specific study or validation study objectives.

Preexisting stress-related diagnoses and mortality: A Danish cancer cohort study.

BACKGROUND: This study evaluated the association between preexisting stress-related diagnoses and mortality in a Danish population-based cancer cohort. METHODS: This study included Danish patients with cancer diagnosed in 1995-2011 who had a stress-related diagnosis before their cancer diagnosis. Cancer patients without a prior stress-related diagnosis were matched 5:1 to the stress disorder cohort by cancer site, age group, calendar period, and sex. The 5-year cumulative incidence of cancer-specific and all-cause mortality was computed by stress-related diagnosis category. Hazard ratios and 95% confidence intervals (CIs) associating stress-related diagnoses with mortality were computed by follow-up time, stress-related diagnosis category, stage, comorbidity status, and cancer type. RESULTS: This study identified 4437 cancer patients with a preexisting stress-related diagnosis and 22,060 matched cancer cohort members. The 5-year cumulative risk of cancer-specific mortality was 33% (95% CI, 32%-35%) for those with a preexisting stress-related diagnosis and 29% (95% CI, 28%-29%) for those without a prior stress-related diagnosis. Cancer patients with a preexisting stress-related diagnosis had a 1.3 times higher cancer-specific mortality rate than the comparison cohort members (95% CI, 1.2-1.5). This increase persisted across categories of stress-related diagnosis. The association varied by stage and cancer type, with more pronounced associations found among those with a late stage at diagnosis and hematological malignancies. CONCLUSIONS: Cancer patients with preexisting stress-related diagnoses had increased rates of cancer-specific and all-cause mortality. The results suggest that psychiatric comorbidities may be an important consideration for cancer prognosis, and cancer treatment informed by a patient's history may improve outcomes.

Drivers of racial, regional, and socioeconomic disparities in late-stage breast cancer mortality.

BACKGROUND: The authors identified tumor, treatment, and patient characteristics that may contribute to differences in breast cancer (BC) mortality by race, rurality, and area-level socioeconomic status (SES) among women diagnosed with stage IIIB-IV BC in Georgia. METHODS: Using the Georgia Cancer Registry, 3084 patients with stage IIIB-IV primary BC (2013-2017) were identified. Cox proportional hazards regression was used to calculate the hazard ratios (HRs) and 95% confidence intervals (CIs) comparing mortality among non-Hispanic Black (NHB) versus non-Hispanic White (NHW), residents of rural versus urban neighborhoods, and residents of low- versus high-SES neighborhoods by tumor, treatment, and patient characteristics. The mediating effects of specific characteristics on the association between race and BC mortality were estimated. RESULTS: Among the study population, 41% were NHB, 21% resided in rural counties, and 72% resided in low SES neighborhoods. The authors observed mortality disparities by race (HR, 1.27; 95% CI, 1.13, 1.41) and rurality (HR, 1.14; 95% CI, 1.00, 1.30), but not by SES (HR, 1.04; 95% CI, 0.91, 1.19). In the stratified analyses, racial disparities were the most pronounced among women with HER2 overexpressing tumors (HR, 2.30; 95% CI, 1.53, 3.45). Residing in a rural county was associated with increased mortality among uninsured women (HR, 2.25; 95% CI, 1.31, 3.86), and the most pronounced SES disparities were among younger women (<40 years: HR, 1.46; 95% CI, 0.88, 2.42). CONCLUSIONS: There is considerable variation in racial, regional, and socioeconomic disparities in late-stage BC mortality by tumor, treatment, and patient characteristics.

Barriers to breast cancer screening in Atlanta, GA: results from the Pink Panel survey at faith-based institutions.

PURPOSE: Our research sought to describe barriers to mammography screening among a sample of predominantly Black women in metropolitan Atlanta, Georgia. METHODS: The Pink Panel project convened community leaders from faith-based institutions to administer an offline survey to women via convenience sampling at fourteen churches in Atlanta in late 2019 and early 2020. With the COVID-19 pandemic, the research team switched to an online survey. The survey included seven questions about breast cancer awareness, barriers to breast cancer screening, and screening status. We used residence information to attain the 9-digit zip code to link to the Area Deprivation Index at the Census Block Group neighborhood level. We report results as descriptive statistics of the barriers to mammography screening. RESULTS: The 643 women represented 21 counties in Georgia, predominantly from metropolitan Atlanta, and 86% identified as Black. Among women aged 40 and older, 90% have ever had a mammogram. Among all women, 79% have ever had a mammogram, and 86% indicated that they would get a mammogram if offered in their neighborhood. The top barriers to mammography screening were lack of health insurance and high cost. Barriers to mammography screening did not differ substantially by Area Deprivation Index. CONCLUSION: Among metropolitan Atlanta women aged 40+ , nearly all reported ever having a mammogram. However, addressing the barriers, including lack of health insurance and high cost, that women reported may further improve mammography screening rates.

Misclassification of Neonatal Abstinence Syndrome Surveillance Estimates: Is Considering the Positive Predictive Value Enough?

BACKGROUND: Validation studies estimating the positive predictive value (PPV) of neonatal abstinence syndrome (NAS) have consistently suggested overreporting in hospital discharge records. However, few studies estimate the negative predictive value (NPV). Even slightly imperfect NPVs have the potential to bias estimated prevalences of rare outcomes like NAS. Given the challenges in estimating NPV, our objective was to evaluate whether the PPV was sufficient to understand the influence of NAS misclassification bias on conclusions of the NAS prevalence in surveillance research. METHODS: We used hospital discharge data from the 2016 New Jersey State Inpatient Databases, Healthcare Cost and Utilization Project. We adjusted surveillance data for misclassification using quantitative bias analysis models to estimate the expected NAS prevalence under a range of PPV and NPV bias scenarios. RESULTS: The 2016 observed NAS prevalence was 0.61%. The misclassification-adjusted prevalence estimates ranged from 0.31% to 0.91%. When PPV was assumed to be ≥90%, the misclassification-adjusted prevalence was typically greater than the observed prevalence but the reverse was true for PPV ≤70%. Under PPV 80%, the misclassification-adjusted prevalence was less than the observed prevalence for NPV >99.9% but flipped for NPV <99.9%. CONCLUSIONS: When we varied the NPV below 100%, our results suggested that the direction of bias (over or underestimation) was dependent on the PPV, and sometimes dependent on the NPV. However, NPV was important for understanding the magnitude of bias. This study serves as an example of how quantitative bias analysis methods can be applied in NAS surveillance to supplement existing validation data when NPV estimates are unavailable.

Time to Surgical Treatment and Facility Characteristics as Potential Drivers of Racial Disparities in Breast Cancer Mortality.

BACKGROUND: Black women are more likely to die of breast cancer than White women. This study evaluated the contribution of time to primary surgical management and surgical facility characteristics to racial disparities in breast cancer mortality among both Black and White women. METHODS: The study identified 2224 Black and 3787 White women with a diagnosis with stages I to III breast cancer (2010-2014). Outcomes included time to surgical treatment (> 30 days from diagnosis) and breast cancer mortality. Odds ratios (ORs) and 95% confidence intervals (CIs) associating surgical facility characteristics with surgical delay were computed, and Cox proportional hazards regression was used to compute hazard ratios (HRs) and 95% CIs associating delay and facility characteristics with breast cancer mortality. RESULTS: Black women were two times more likely to have a surgical delay (OR, 2.15; 95% CI, 1.92-2.41) than White women. Racial disparity in surgical delay was least pronounced among women treated at a non-profit facility (OR, 1.95; 95% CI, 1.70-2.25). The estimated mortality rate for Black women was two times that for White women (HR, 2.00; 95% CI, 1.83-2.46). Racial disparities in breast cancer mortality were least pronounced among women who experienced no surgical delay (HR, 1.81; 95% CI, 1.28-2.56), received surgery at a government facility (HR, 1.31; 95% CI, 0.76-2.27), or underwent treatment at a Commission on Cancer-accredited facility (HR, 1.82; 95% CI, 1.38-2.40). CONCLUSIONS: Black women were more likely to experience a surgical delay and breast cancer death. Persistent racial disparities in breast cancer mortality were observed across facility characteristics except for government facilities.

Hypoxia-inducible factor-1α expression and breast cancer recurrence in a Danish population-based case control study.

BACKGROUND: Hypoxia-inducible factor-1α (HIF-1α) is a transcription factor that facilitates the adaptation of cancer cells to hypoxic conditions and may be prognostic of breast cancer recurrence. We evaluated the association of HIF-1α expression with breast cancer recurrence, and its association with timing of breast cancer recurrence. METHODS: In this population-based case-control study, we included women diagnosed with stage I-III breast cancer between 1985 and 2001, aged 35-69 years, registered in the Danish Breast Cancer Group. We identified 541 cases of breast cancer recurrence among women with estrogen receptor (ER)-positive disease who were treated with tamoxifen for at least 1 year (ER+ TAM+). We also enrolled 300 breast cancer recurrence cases among women with ER-negative disease, not treated with tamoxifen, who survived at least 1 year (ER-/TAM-). Controls were recurrence-free breast cancer patients at the time of case diagnosis, matched to recurrence cases on ER/TAM status, date of surgery, menopausal status, cancer stage, and county of residence. Expression of HIF-1α was measured by immunohistochemistry on tissue microarrays. We fitted logistic regression models to compute odds ratios (ORs) and 95% confidence intervals (CIs) associating HIF-1α expression with recurrence, and with timing of recurrence. RESULTS: HIF-1α expression was observed in 23% of cases and 20% of controls in the ER+/TAM+ stratum, and in 47% of cases and 48% of controls in the ER-/TAM- stratum. We observed a near-null association between HIF-1α expression in both ER/TAM groups (ER+/TAM+ OR = 1.21, 95%CI 0.88, 1.67 and ER-/TAM- OR = 0.97, 95%CI 0.68, 1.39). HIF-1α expression was not associated with time to recurrence among women in the ER+/TAM+ stratum, but was associated with early recurrence among women in the ER-/TAM- stratum. CONCLUSION: In this study, HIF-1α expression was not associated with breast cancer recurrence overall but may be associated with early recurrence among women diagnosed with ER- breast cancer.

Bias Analysis Gone Bad.

Quantitative bias analysis comprises the tools used to estimate the direction, magnitude, and uncertainty from systematic errors affecting epidemiologic research. Despite the availability of methods and tools, and guidance for good practices, few reports of epidemiologic research incorporate quantitative estimates of bias impacts. The lack of familiarity with bias analysis allows for the possibility of misuse, which is likely most often unintentional but could occasionally include intentional efforts to mislead. We identified 3 examples of suboptimal bias analysis, one for each common bias. For each, we describe the original research and its bias analysis, compare the bias analysis with good practices, and describe how the bias analysis and research findings might have been improved. We assert no motive to the suboptimal bias analysis by the original authors. Common shortcomings in the examples were lack of a clear bias model, computed example, and computing code; poor selection of the values assigned to the bias model's parameters; and little effort to understand the range of uncertainty associated with the bias. Until bias analysis becomes more common, community expectations for the presentation, explanation, and interpretation of bias analyses will remain unstable. Attention to good practices should improve quality, avoid errors, and discourage manipulation.

Estimating the Unknown: Greater Racial and Ethnic Disparities in COVID-19 Burden After Accounting for Missing Race and Ethnicity Data.

BACKGROUND: Black, Hispanic, and Indigenous persons in the United States have an increased risk of SARS-CoV-2 infection and death from COVID-19, due to persistent social inequities. However, the magnitude of the disparity is unclear because race/ethnicity information is often missing in surveillance data. METHODS: We quantified the burden of SARS-CoV-2 notification, hospitalization, and case fatality rates in an urban county by racial/ethnic group using combined race/ethnicity imputation and quantitative bias analysis for misclassification. RESULTS: The ratio of the absolute racial/ethnic disparity in notification rates after bias adjustment, compared with the complete case analysis, increased 1.3-fold for persons classified Black and 1.6-fold for those classified Hispanic, in reference to classified White persons. CONCLUSIONS: These results highlight that complete case analyses may underestimate absolute disparities in notification rates. Complete reporting of race/ethnicity information is necessary for health equity. When data are missing, quantitative bias analysis methods may improve estimates of racial/ethnic disparities in the COVID-19 burden.

Receipt of Guideline-Concordant Care Does Not Explain Breast Cancer Mortality Disparities by Race in Metropolitan Atlanta.

BACKGROUND: Racial disparities in breast cancer mortality in the United States are well documented. Non-Hispanic Black (NHB) women are more likely to die of their disease than their non-Hispanic White (NHW) counterparts. The disparity is most pronounced among women diagnosed with prognostically favorable tumors, which may result in part from variations in their receipt of guideline care. In this study, we sought to estimate the effect of guideline-concordant care (GCC) on prognosis, and to evaluate whether receipt of GCC modified racial disparities in breast cancer mortality. PATIENTS AND METHODS: Using the Georgia Cancer Registry, we identified 2,784 NHB and 4,262 NHW women diagnosed with a stage I-III first primary breast cancer in the metropolitan Atlanta area, Georgia, between 2010 and 2014. Women were included if they received surgery and information on their breast tumor characteristics was available; all others were excluded. Receipt of recommended therapies (chemotherapy, radiotherapy, endocrine therapy, and anti-HER2 therapy) as indicated was considered GCC. We used Cox proportional hazards models to estimate the impact of receiving GCC on breast cancer mortality overall and by race, with multivariable adjusted hazard ratios (HRs). RESULTS: We found that NHB and NHW women were almost equally likely to receive GCC (65% vs 63%, respectively). Failure to receive GCC was associated with an increase in the hazard of breast cancer mortality (HR, 1.74; 95% CI, 1.37-2.20). However, racial disparities in breast cancer mortality persisted despite whether GCC was received (HRGCC: 2.17 [95% CI, 1.61-2.92]; HRnon-GCC: 1.81 [95% CI, 1.28-2.91] ). CONCLUSIONS: Although receipt of GCC is important for breast cancer outcomes, racial disparities in breast cancer mortality did not diminish with receipt of GCC; differences in mortality between Black and White patients persisted across the strata of GCC.

Bayesian Pathway Analysis for Complex Interactions.

Modern epidemiologic studies permit investigation of the complex pathways that mediate effects of social, behavioral, and molecular factors on health outcomes. Conventional analytical approaches struggle with high-dimensional data, leading to high likelihoods of both false-positive and false-negative inferences. Herein, we describe a novel Bayesian pathway analysis approach, the algorithm for learning pathway structure (ALPS), which addresses key limitations in existing approaches to complex data analysis. ALPS uses prior information about pathways in concert with empirical data to identify and quantify complex interactions within networks of factors that mediate an association between an exposure and an outcome. We illustrate ALPS through application to a complex gene-drug interaction analysis in the Predictors of Breast Cancer Recurrence (ProBe CaRe) Study, a Danish cohort study of premenopausal breast cancer patients (2002-2011), for which conventional analyses severely limit the quality of inference.

Adaptive Validation Design: A Bayesian Approach to Validation Substudy Design With Prospective Data Collection.

An internal validation substudy compares an imperfect measurement of a variable with a gold-standard measurement in a subset of the study population. Validation data permit calculation of a bias-adjusted estimate, which has the same expected value as the association that would have been observed had the gold-standard measurement been available for the entire study population. Existing guidance on optimal sampling for validation substudies assumes complete enrollment and follow-up of the target cohort. No guidance exists for validation substudy design while cohort data are actively being collected. In this article, we use the framework of Bayesian monitoring methods to develop an adaptive approach to validation study design. This method monitors whether sufficient validation data have been collected to meet predefined criteria for estimation of the positive and negative predictive values. We demonstrate the utility of this method using the Study of Transition, Outcomes and Gender-a cohort study of transgender and gender nonconforming people. We demonstrate the method's ability to determine efficacy (when sufficient validation data have accumulated to obtain estimates of the predictive values that fall above a threshold value) and futility (when sufficient validation data have accumulated to conclude the mismeasured variable is an untenable substitute for the gold-standard measurement). This proposed method can be applied within the context of any parent epidemiologic study design and modified to meet alternative criteria given specific study or validation study objectives. Our method provides a novel approach to effective and efficient estimation of classification parameters as validation data accrue.