Search

To assess a new approach (weighting by "income probabilities [IP]") that uses US Census data from the patients' communities to approximate individual-level income, an important but often missing variable in health services research.Community (census tract level) income data came from the 2017 5-year American Community Survey (ACS). The patient data included those diagnosed with cancer in 2017 in Ohio (n = 65,759). The reference population was the 2017 5-year ACS Public Use Microdata Sample (n = 564,357 generalizing to 11,288,350 Ohioans).We applied the traditional approach of income approximation using median census tract income along with two IP based approaches to estimate the proportions in the patient data with incomes of 0%-149%, 150%-299%, 300%-499%, and 500%+ of the federal poverty level (FPL) ("class-relevant income grouping") or 0%-138%, 139%-249%, 250%-399%, and 400%+ FPL ("policy-relevant income grouping"). These estimated income distributions were then compared with the known income distributions of the reference population.The patient data came from Ohio's cancer registry. The other data were publicly available.Both IP based approaches consistently outperformed the traditional approach overall and in subgroup analyses, as measured by the weighted average absolute percentage point differences between the proportions of each of the income categories of the reference population and the estimated proportions generated by the income approximation approaches ("average percent difference," or APD). The smallest APD for an IP based method, 0.5%, was seen in non-Hispanic White females in the class-relevant income grouping (compared with 16.5% for the conventional method), while the largest APD, 7.1%, was seen in non-Hispanic Black females in the policy-relevant income grouping (compared with 18.0% for the conventional method).Weighting by IP substantially outperformed the conventional approach of estimating the distribution of incomes in patient data.

CONTEXT: Prior studies demonstrate that Medicaid expansion has been associated with earlier stage breast cancer diagnosis among women with low-income, likely through increased access to cancer screening services. However, how this policy change has impacted geospatial disparities in breast cancer stage at diagnosis is unclear. OBJECTIVE: To examine whether there were reductions in geospatial disparities in advanced stage breast cancer at diagnosis in Ohio after Medicaid expansion. DESIGN: The study included 33 537 women aged 40-64 years diagnosed with invasive breast cancer from the Ohio Cancer Incidence Surveillance System between 2010 and 2017. The space-time scan statistic was used to detect clusters of advanced stage at diagnosis before and after Medicaid expansion. Block group variables from the Census were used to describe the contextual characteristics of detected clusters. RESULTS: The percentage of local stage diagnosis among women with breast cancer increased from 60.2% in the pre-expansion period (2010-2013) to 62.6% in the post-expansion period (2014-2017), while the uninsured rate among those women decreased from 13.7% to 7.5% during the same period. Two statistically significant (p < .05) and six non-significant spatial clusters (p > .05) of advanced stage breast cancer cases were found in the pre-expansion period, while none were found in the post-expansion period. These clusters were in the four largest metropolitan areas in Ohio and individuals inside the clusters were more likely to be disadvantaged along numerous socioeconomic factors. CONCLUSIONS: Medicaid expansion has played an important role in reducing geospatial disparities in breast cancer stage at diagnosis, likely through the reduction of advanced stage disease among women living in socioeconomically disadvantaged communities.

PurposeA disconnect often exists between those with the expertise to manage and analyze complex, multi-source data sets, and the clinical, social services, advocacy, and public health professionals who can pose the most relevant questions and best apply the answers. We describe development and implementation of a cancer informatics infrastructure aimed at broadening the usability of community cancer data to inform cancer control research and practice; and we share lessons learned. MethodsWe built a multi-level database known as The Ohio Cancer Assessment and Surveillance Engine (OH-CASE) to link data from Ohio’s cancer registry with community data from the U.S. Census and other sources. Space- and place-based characteristics were assigned to individuals according to residential address. Stakeholder input informed development of an interface for generating queries based on geographic, demographic, and disease inputs and for outputting results aggregated at the state, county, municipality, or zip code levels. ResultsOH-CASE contains 723,410 patient records for Ohioans diagnosed with cancer from 1/1/2006 – 12/31/2017 across 88 counties containing 1215 municipalities and 1197 zip codes. Stakeholder feedback from cancer center community outreach teams, advocacy organizations, public health, and researchers suggests a broad range of uses of such multi-level data resources accessible via a user interface. ConclusionOH-CASE represents a prototype of a transportable model for curating and synthesizing data to understand cancer burden across communities. Beyond supporting collaborative research, this infrastructure can serve the clinical, social services, public health, and advocacy communities by enabling targeting of outreach, funding, and interventions to narrow cancer disparities.

Background Older patients with myelodysplastic syndromes (MDS), particularly those with no or one cytopenia and no transfusion dependence, typically have an indolent course. Approximately, half of these receive the recommended diagnostic evaluation (DE) for MDS. We explored factors determining DE in these patients and its impact on subsequent treatment and outcomes. Patients and Methods We used 2011-2014 Medicare data to identify patients ≥66 years of age diagnosed with MDS. We used Classification and Regression Tree (CART) analysis to identify combinations of factors associated with DE and its impact on subsequent treatment. Variables examined included demographics, comorbidities, nursing home status, and investigative procedures performed. We conducted a logistic regression analysis to identify correlates associated with receipt of DE and treatment. Results Of 16 851 patients with MDS, 51% underwent DE. patients with MDS with no cytopenia (n = 3908) had the lowest uptake of DE (34.7%). Compared to patients with no cytopenia, those with any cytopenia had nearly 3 times higher odds of receiving DE [adjusted odds ratio (AOR), 2.81: 95% CI, 2.60-3.04] and the odds were higher for men than for women [AOR, 1.39: 95%CI, 1.30-1.48] and for Non-Hispanic Whites [vs. everyone else (AOR, 1.17: 95% CI, 1.06-1.29)]. The CART showed DE as the principal discriminating node, followed by the presence of any cytopenia for receiving MDS treatment. The lowest percentage of treatment was observed in patients without DE, at 14.6%. Conclusion In this select older patients with MDS, we identified disparities in accurate diagnosis by demographic and clinical factors. Receipt of DE influenced subsequent treatment but not survival.

Background: Prior studies have shown significant improvements in earlier breast cancer stage at diagnosis following Medicaid expansion. However, the mechanisms behind this finding have not been explored. The timing of Medicaid enrollment among cancer patients has been shown to be associated with cancer stage at diagnosis, with significantly higher percentages of women with late-stage breast cancer enrolling in Medicaid shortly before or after cancer diagnosis (hereafter referred to as the Peri-dx group), compared to those continuously enrolled in Medicaid at least 4 months prior to cancer diagnosis (Pre-dx group). Hence, we conducted a causal mediation study based on a logistic regression model to determine whether the observed improvements in breast cancer stage at diagnosis from Pre- to Post-expansion can be attributed to a decreasing percentage of women in the Peri-dx group, and a corresponding increase in the Pre-dx group. Methods: We used the linked Ohio Cancer Incidence Surveillance System and Medicaid data, and identified women in Ohio age younger than 66 years, diagnosed with incident invasive breast cancer during the period May 2011-December 2017, and enrolled in Medicaid at the time of cancer diagnosis (n=5,880, excluding those with unstaged/unknown-stage cancer). We defined the years 2011-2013 and 2014-2017 as the Pre- and Post-expansion periods, and Month 0 as the month of cancer diagnosis. We identified women enrolled in Medicaid in the -3 to +3-month window in the Peri-dx group, and those continuously enrolled in Medicaid at least 4 months prior to cancer diagnosis in the Pre-dx group. We conducted a causal mediation analysis to estimate the direct, indirect, and total effect of Medicaid expansion on being diagnosed with local-stage disease versus late- (regional- and distant-) stage disease, with the mediator variable being a patient in the Pre-dx group (versus in the Peri-dx group), adjusting for patient- and area-level covariates. Results: The median age of our study population was 53 years; 75.0% were non-Hispanic White. The percent of breast cancer patients diagnosed with late-stage disease was 43.5% and 53.4% in the Pre- and Peri-dx groups, respectively. We observed a clear shift in the distribution of breast cancer patients from the Peri- to the Pre-dx group over time, with an increase in the percentage of women in the Pre-dx group (from 55.2% to 74.2% from the Pre- to Post-Expansion period), and a corresponding decrease in the Peri-dx group (from 44.8% to 25.8%). Findings from our causal mediation analysis, based on a risk difference scale, showed that while the direct effect of Medicaid expansion on being diagnosed with local-stage disease was not significant (coefficient: 0.037, (p=0.297)), both the indirect and total effects were highly significant (0.049 (p < 0.001), and 0.086 (p=0.015)). Conclusion: Our results indicate that the improvements in breast cancer stage at diagnosis from Pre- to Post-Expansion can be attributed to a decrease in the Peri-dx group, and a corresponding increase in the Pre-dx group, which improves access to screening, increases the likelihood of breast cancer patients being engaged with Medicaid providers, and reduces delays in enrollment logistics prior to treatment initiation. Citation Format: Siran M. Koroukian, Weichuan Dong, Jeffrey Albert, Uriel Kim, Johnie Rose, Cynthia Owusu, Kristine Zanotti, Gregory Cooper, Jennifer Tsui. Impact of medicaid expansion on breast cancer stage at diagnosis: Exploring the mechanisms at play [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P3-12-30.

BACKGROUND. The objective of this study is to understand the effect of Medicaid expansion under the Affordable Care Act (ACA) on patterns of surgical care among low-income breast cancer patients. Emerging literature suggests cancer patients in Medicaid expansion states are presenting with earlier stages of disease. However, less is known regarding the implications of Medicaid expansion on patterns of surgical care in low-income women. PATIENTS AND METHODS. We compared nonmetastatic 30–64-year-old uninsured or Medicaid-insured Ohio breast cancer patients diagnosed 4 years before and 4 years after the state’s 2014 Medicaid expansion (study group); the control group was the privately insured. Time-to-surgery (TTS) was defined as days from diagnosis to surgery. Demographic and treatment variables before and after expansion were examined in multivariate analysis. RESULTS. There was a 10.4% point increase in breast conservation therapy (BCT) in the study group (pre-ACA 26.3%, post-ACA 36.7%; p < 0.01) compared with a 5.8% point increase in the control group (pre-ACA 36.0%, post-ACA 41.8%; p < 0.01). Disparities in reconstruction narrowed between the study (pre-ACA 21.4%, post-ACA 34.5%; p < 0.01) and the control (37.0% pre-ACA, 44.1% post-ACA group p < 0.01) groups. There was no statistically significant change in mean TTS in the study group (pre-ACA 42.1 days, post-ACA 43.1 days p = 0.18) but there was an increase in TTS in the control group (pre-ACA 35.0 days, post ACA 37.0 days; p < 0.01). CONCLUSIONS. Medicaid expansion appears to have narrowed disparities in the utilization of BCT and reconstruction in low-income women. However, there was no improvement in surgical delay.

The mechanisms underlying improvements in early-stage cancer at diagnosis following Medicaid expansion remain unknown. We hypothesized that Medicaid expansion allowed for low-income adults to enroll in Medicaid before cancer diagnosis, thus increasing the number of stably-enrolled relative to those who enroll in Medicaid only after diagnosis (emergently-enrolled).Using data from the 2011-2017 Ohio Cancer Incidence Surveillance System and Medicaid enrollment files, we identified individuals diagnosed with incident invasive breast (n=4850), cervical (n=1023), and colorectal (n=3363) cancer. We conducted causal mediation analysis to estimate the direct effect of pre- (vs. post-) expansion on being diagnosed with early-stage (-vs. regional-stage and distant-stage) disease, and indirect (mediation) effect through being in the stably- (vs. emergently-) enrolled group, controlling for individual-level and area-level characteristics.The percentage of stably-enrolled patients increased from 63.3% to 73.9% post-expansion, while that of the emergently-enrolled decreased from 36.7% to 26.1%. The percentage of patients with early-stage diagnosis remained 1.3-2.9 times higher among the stably-than the emergently-enrolled group, both pre-expansion and post-expansion. Results from the causal mediation analysis showed that there was an indirect effect of Medicaid expansion through being in the stably- (vs. emergently-) enrolled group [risk ratios with 95% confidence interval: 1.018 (1.010-1.027) for breast cancer, 1.115 (1.064-1.167) for cervical cancer, and 1.090 (1.062-1.118) for colorectal cancer.We provide the first evidence that post-expansion improvements in cancer stage were caused by an increased reliance on Medicaid as a source of stable insurance coverage.

ImportanceThe association between cancer mortality and risk factors may vary by geography. However, conventional methodological approaches rarely account for this variation.ObjectiveTo identify geographic variations in the association between risk factors and cancer mortality.Design, Setting, and ParticipantsThis geospatial cross-sectional study used county-level data from the National Center for Health Statistics for individuals who died of cancer from 2008 to 2019. Risk factor data were obtained from County Health Rankings & Roadmaps, Health Resources and Services Administration, and Centers for Disease Control and Prevention. Analyses were conducted from October 2021 to July 2022.Main Outcomes and MeasuresConventional random forest models were applied nationwide and by US region, and the geographical random forest model (accounting for local variation of association) was applied to assess associations between a wide range of risk factors and cancer mortality.ResultsThe study included 7 179 201 individuals (median age, 70-74 years; 3 409 508 women [47.5%]) who died from cancer in 3108 contiguous US counties during 2008 to 2019. The mean (SD) county-level cancer mortality rate was 177.0 (26.4) deaths per 100 000 people. On the basis of the variable importance measure, the random forest models identified multiple risk factors associated with cancer mortality, including smoking, receipt of Supplemental Nutrition Assistance Program (SNAP) benefits, and obesity. The geographical random forest model further identified risk factors that varied at the county level. For example, receipt of SNAP benefits was a high-importance factor in the Appalachian region, North and South Dakota, and Northern California; smoking was of high importance in Kentucky and Tennessee; and female-headed households were high-importance factors in North and South Dakota. Geographic areas with certain high-importance risk factors did not consistently have a corresponding high prevalence of the same risk factors.Conclusions and RelevanceIn this cross-sectional study, the associations between cancer mortality and risk factors varied by geography in a way that did not correspond strictly to risk factor prevalence. The degree to which other place-specific characteristics, observed and unobserved, modify risk factor effects should be further explored, and this work suggests that risk factor importance may be a preferable paradigm for selecting cancer control interventions compared with risk factor prevalence.

Disparities in premature cardiovascular mortality (PCVM) have been associated with socioeconomic, behavioral, and environmental risk factors. Understanding the “phenotypes”, or combinations of characteristics associated with the highest risk of PCVM, and the geographic distributions of these phenotypes, is critical to targeting PCVM interventions. This study applied classification and regression tree (CART) to identify county phenotypes of PCVM and geographic information systems to examine the distributions of identified phenotypes. Random forest analysis was applied to evaluate relative importance of risk factors associated with PCVM. The CART analysis identified seven county phenotypes of PCVM, where high-risk phenotypes were characterized as having greater percentages of people with lower income, higher physical inactivity, and higher food insecurity. These high-risk phenotypes were mostly concentrated in the Black Belt of the American South and the Appalachian region. The random forest analysis identified additional risk factors associated with PCVM, including broadband access, smoking, receipt of Supplemental Nutrition Assistance Program benefits, excessive drinking, and educational attainment. Overall, county-level risk factors explained 52% of the inter-county variation in PCVM. Our study demonstrates the use of machine learning approaches in characterizing area-level phenotypes of PCVM. Interventions to reduce PCVM should be tailored according to these phenotypes in corresponding geographic areas.

Introduction: Substantial delays in time to initiation of treatment (TTT) following diagnosis of breast cancer (BC) can inflict a toll on quality of life and can decrease cancer-specific survival. Among low-income, non-elderly Ohio women having newly-diagnosed breast cancer with local or regional spread, we examined TTT and other measures where income-related disparities have been documented, comparing 2011-2013 (pre-Medicaid-expansion) vs. 2014-2016 (post-expansion). Our primary hypothesis was that TTT would decrease following 2014 Medicaid expansion.Methods: Using data from the Northeast Ohio Cancer Assessment and Surveillance Engine (NEO-CASE), a multilevel data infrastructure linking Ohio cancer registry data with community data, we identified 30-64 year-old women with new diagnosis of invasive, non-metastatic BC who were uninsured or on Medicaid when diagnosed. TTT was defined as days from diagnosis to first BC treatment with any modality. We excluded women with TTT=0 (likely coding error). The main exposure was pre- or post-Medicaid expansion period defined as 2011-2013 or 2014-2016, respectively. We examined additional key demographic and treatment variables before and after expansion and in multivariate analysis of TTT. We used a previously-described probability-weighting approach based on neighborhood median income to approximate excluding women with incomes above 138% of Federal Poverty Level. As a control analysis, we compared pre- and post-expansion TTT among privately insured women, probability weighted to select for higher income individuals.Results: Our study population included 1177 and 1143 women diagnosed with BC pre- and post-expansion, respectively. Demographic characteristics were similar, though mean age increased by 1.2 years (p Pre-expansionPost-expansionpN11771143Mean TTT in days (SD)41.1 (37.1)43.1 (33.7)0.18Mean age at diagnosis (SD)51.6 (8.39)52.8 (8.14) Citation Format: Johnie Rose, Weichuan Dong, Uriel Kim, Samilia Obeng-Gyasi, Siran Koroukian. Medicaid expansion associated with earlier stage and improved reconstruction rates in low income breast cancer patients [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr SS1-09.

In recent years, deep learning has been widely used in hyperspectral image (HSI) classification and has shown good capabilities. Particularly, the use of convolutional neural network (CNN) in HSI classification has achieved attractive performance. However, HSI contains a lot of redundant information, and the CNN-based model is limited by the receptive field of CNN and cannot balance the performance and depth of the model. Furthermore, considering that HSI can be regarded as sequence data, CNN-based models cannot mine sequence features well. In this paper, we propose a model named SSA-Transformer to address the above problems and extract spectral-spatial features of HSI more efficiently. The SSA-Transformer model combines a modified CNN-based spectral-spatial attention mechanism and a self-attention-based transformer with dense connection. The SSA-Transformer model can combine the local and global features of HSI to improve the performance of the model. A series of experiments showed that the SSA-Transformer achieved competitive classification accuracy compared with other CNN-based classification methods using three HSI datasets: University of Pavia (PU), Salinas (SA), and Kennedy Space Center (KSC).

The proportion of patients diagnosed with colorectal cancer (CRC) at age < 50 (early-onset CRC, or EOCRC) has steadily increased over the past three decades relative to the proportion of patients diagnosed at age ≥ 50 (late-onset CRC, or LOCRC), despite the reduction in CRC incidence overall. An important gap in the literature is whether EOCRC shares the same community-level risk factors as LOCRC. Thus, we sought to (1) identify disparities in the incidence rates of EOCRC and LOCRC using geospatial analysis and (2) compare the importance of community-level risk factors (racial/ethnic, health status, behavioral, clinical care, physical environmental, and socioeconomic status risk factors) in the prediction of EOCRC and LOCRC incidence rates using a random forest machine learning approach. The incidence data came from the Surveillance, Epidemiology, and End Results program (years 2000–2019). The geospatial analysis revealed large geographic variations in EOCRC and LOCRC incidence rates. For example, some regions had relatively low LOCRC and high EOCRC rates (e.g., Georgia and eastern Texas) while others had relatively high LOCRC and low EOCRC rates (e.g., Iowa and New Jersey). The random forest analysis revealed that the importance of community-level risk factors most predictive of EOCRC versus LOCRC incidence rates differed meaningfully. For example, diabetes prevalence was the most important risk factor in predicting EOCRC incidence rate, but it was a less important risk factor of LOCRC incidence rate; physical inactivity was the most important risk factor in predicting LOCRC incidence rate, but it was the fourth most important predictor for EOCRC incidence rate. Thus, our community-level analysis demonstrates the geographic variation in EOCRC burden and the distinctive set of risk factors most predictive of EOCRC.

Background: Medicaid expansion in Ohio went into effect in 2014. While prior studies have shown earlier breast cancer stages at diagnosis post-expansion, its impact relative to receipt of standard treatment remains understudied. Here, we examined changes in stage-appropriate standard treatment by women with invasive incident breast cancer pre- and post-expansion, and by their timing of enrollment in Medicaid relative to cancer diagnosis – a key marker of inadequate access to care. Methods: We used 2011-2017 Ohio Cancer Incidence Surveillance System (OCISS) data linked with Medicaid enrollment and claims files. Our study population included women 18-65 years of age diagnosed with local- or regional-stage cancer (n=1,053 and n=2,264 women in the pre- and post-expansion periods, respectively). We identified women as Previously Enrolled if they enrolled in Medicaid at least 3 months prior to cancer diagnosis and remained continuously enrolled for 4 months post-diagnosis, and as Newly Enrolled if they enrolled in Medicaid in the 3 months before or after cancer diagnosis. We used the OCISS for sociodemographic and treatment data. We defined standard treatment as the receipt of mastectomy, or lumpectomy and radiation in all women. In those with regional-stage cancer, we required standard treatment to also include chemotherapy; and for those with positive hormonal status, we required hormonal therapy. From claims data, we identified comorbid conditions, grouped as none, physical conditions only, mental conditions only, and physical and mental conditions, with or without substance abuse. In addition to descriptive analysis, we conducted multivariable logistic regression analysis to evaluate the independent association between Medicaid expansion and receipt of standard treatment, after adjusting for patient demographics, tumor characteristics, and comorbid conditions. Results: The median age was 50.6 years and 51.8 years in the pre- and post-expansion periods. The percent of Newly Enrolled women relative to the total decreased from 24.8% to 18.9%. Women in this group were healthier than those in the Previously Enrolled group, as evidenced by the higher percent of women with no comorbid conditions (43.3% and 42.4% vs. 21.7% and 22.5%, Pre- and Post-expansion, respectively). From Pre- to Post-Expansion, receipt of standard treatment increased from 52.3% to 58.8% and from 58.6% to 61.6% among the Previously and Newly Enrolled, respectively. Results from the multivariable logistic regression analysis showed that compared to women diagnosed in the pre-expansion period, those diagnosed in the post-expansion period had 10% higher odds to receive standard treatment, after adjusting for patient covariates. Conversely, the timing of enrollment in Medicaid relative to cancer diagnosis was not associated with our outcome of interest. Conclusion: Medicaid expansion in Ohio was associated with improvements in receipt of stage-appropriate standard treatment for breast cancer. Future studies should elucidate the mechanisms at play. Citation Format: Siran M. Koroukian, Weichuan Dong, Kirsten Eom, Uriel M. Kim, Johnie Rose, Jeffrey Albert, Cynthia Owusu, Kristine Zanotti, Gregory S. Cooper, Jennifer Tsui. Receipt of standard treatment by breast cancer patients on Medicaid, pre- and post-expansion [abstract]. In: Proceedings of the 15th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2022 Sep 16-19; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2022;31(1 Suppl):Abstract nr A061.

Objective: Recent data suggest low use of HMAs in patients with MDS in real world. We sought to identify factors associated with HMA use. Methods: Using 2011-2014 Medicare data, we identified individuals ³65 years of age diagnosed with incident MDS during the years 2012-2013 and receiving their care exclusively through the Medicare fee-for-service system. Collected data included age, race, sex, morbidity burden based on their frailty status (robust, pre-frail, mildly frail, moderately, or severely frail), nursing home residence, presence of MDS-related clinical characteristics [cytopenia (0-1 or ³2) and transfusion dependence (independent, low, or medium/high)], area-level socioeconomic measures [percent of individuals with income 100% below the federal poverty level; percent of adults (age > 25 years) without high school diploma; rurality, based on the rural-urban continuum code; and availability of general internal medicine subspecialists per 100,000 population], and receipt of bone marrow biopsy. Among those who received HMA, we grouped patients by the number of complete cycles (1-2, 3-5, or ³6). We used multivariable logistic regression analysis to evaluate the association between demographic characteristics and receipt of HMA, adjusting for MDS-related clinical characteristics, frailty, and socioeconomic measures. Results: Of 49,154 MDS pts, only 16.1% (7,935) received HMA. This percentage ranged from a low of 7.4% among individuals ³85 years of age to a high of 23.2% among those in the 65-74 age group. Receipt of HMA was significantly higher in men than in women (19.0% vs. 12.8%), in Whites than in minoritized patients (16.7% vs. 11.6% in Blacks, and 12.7% in those of Other race), higher among those with 2-3 cytopenias vs. 0-1 cytopenia and highest among those with medium/high transfusion dependence (44.7%). Among those who underwent biopsy, receipt of HMA was at 22.9%. Patients with moderate or severe frailty, and those in a nursing home had significantly lower rates of receipt of HMA. After adjusting for MDS-related clinical variables, frailty and socioeconomic measures, older patients had significantly lower odds than their younger counterparts to receive HMA [adjusted odds ratio (aOR): 0.84, (95% confidence interval: 0.79-0.89), and aOR: 0.43 (0.40-0.47) for those in the 75-84 and ³85 age groups, respectively). Females and minoritized patients had significantly lower odds than men and White patients to receive HMA [(aOR: 0.83 (0.79-0.88) for females; aOR: 0.71 (0.63-0.81) for Blacks, and aOR: 0.78 (0.68-0.89) for patients of Other race). Among those receiving HMA, we did not observe differences in the number of completed cycles by age, race, or sex. Conclusions: In Medicare beneficiaries with MDS, older patients, women, and minoritized patients have significantly lower odds to be treated with HMAs, even after adjusting for MDS-related clinical variables, morbidity burden, and socioeconomic measures. In addition to barriers to access adequate care, these results clearly reflect biases in MDS treatment, favoring younger, White men. Citation Format: Sudipto Mukherjee, Weichuan Dong, Aaron T. Gerds, Hetty E. Carraway, Abhay Singh, Anjali S. Advani, Siran M. Koroukian. Age-, sex-, and race-related disparities in receipt of HypomethyLating Agents (HMA) for the treatment of myelodysplastic syndrome (MDS) [abstract]. In: Proceedings of the 15th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2022 Sep 16-19; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2022;31(1 Suppl):Abstract nr A062.

BackgroundBlack women diagnosed with breast cancer in the U.S. tend to experience significantly longer waits to begin treatment than do their white counterparts, and such treatment delay has been associated with poorer survival. We sought to identify the factors driving or mitigating treatment delay among Black women in an urban community where treatment delay is common.MethodsApplying the SaTScan method to data from Ohio’s state cancer registry, we identified the community within Cuyahoga County, Ohio (home to Cleveland) with the highest degree of breast cancer treatment delay from 2010 through 2015. We then recruited breast cancer survivors living in the target community, their family caregivers, and professionals serving breast cancer patients in this community. Participants completed semi-structured interviews focused on identifying barriers to and facilitators of timely breast cancer treatment initiation after diagnosis.ResultsFactors reported to impact timely treatment fell into three primary themes:informational,intrapersonal, andlogistical. Informational barriers included erroneous beliefs and lack of information about processes of care; intrapersonal barriers centered on mistrust, fear, and denial; while logistical barriers involved transportation and financial access, as well as patients’ own caregiving obligations. An informational facilitator was the provision of objective and understandable disease information, and a common intrapersonal facilitator was faith. Logistical facilitators included financial counseling and mechanisms to assist with Medicaid enrollment. Crosscutting these themes, and mentioned frequently, was the centrality of both patient navigators and support networks (formal and, especially, informal) as critical lifelines for overcoming barriers and leveraging facilitating factors.ConclusionsThe present study describes the numerous hurdles to timely breast cancer treatment faced by Black women in a high-risk urban community. These hurdles, as well as corresponding facilitators, can be classified as informational, intrapersonal, and logistical. Observing similar results on a larger scale could inform the design of interventions and policies to reduce race-based disparities in processes of cancer care.

Background: Disparities in the stage at diagnosis for breast cancer have been independently associated with various contextual characteristics. Understanding which combinations of these characteristics indicate highest risk, and where they are located, is critical to targeting interventions and improving outcomes for patients with breast cancer. Methods: The study included women diagnosed with invasive breast cancer between 2009 and 2018 from 680 U.S. counties participating in the Surveillance, Epidemiology, and End Results program. We used a machine learning approach called Classification and Regression Tree (CART) to identify county “phenotypes,” combinations of characteristics that predict the percentage of patients with breast cancer presenting with late-stage disease. We then mapped the phenotypes and compared their geographic distributions. These findings were further validated using an alternate machine learning approach called random forest. Results: We discovered seven phenotypes of late-stage breast cancer. Common to most phenotypes associated with high risk of late-stage diagnosis were high uninsured rate, low mammography use, high area deprivation, rurality, and high poverty. Geographically, these phenotypes were most prevalent in southern and western states, while phenotypes associated with lower percentages of late-stage diagnosis were most prevalent in the northeastern states and select metropolitan areas. Conclusions: The use of machine learning methods of CART and random forest together with geographic methods offers a promising avenue for future disparities research. Impact: Local interventions to reduce late-stage breast cancer diagnosis, such as community education and outreach programs, can use machine learning and geographic modeling approaches to tailor strategies for early detection and resource allocation.

Designing a system to solve school bus routing problems (SBRP), especially in a large school district, is very complex and expensive. One of the challenges resides in designing routes for the school buses when they have mixed loads, where each bus transports students for one or more schools at the same time to save the total number of buses required. This article aims to explore whether the algorithm originally developed for pickup and delivery problem with time windows (PDPTW) can be employed for solving mixed load SBRP. We present a PDPTW-based algorithm to address the mixed load SBRP focusing on minimizing the number of buses required. Our algorithm combines a record-to-record travel framework with three neighborhood operators, single paired insertion, swapping pairs between routes, and within route insertion, to improve solution iteratively. Results from implementing this algorithm show that our PDPTW-based algorithm is feasible for mixed load SBRP. Moreover, we found that guided strategy is better than random for permuting nodes that would be selected to relocate positions in a route or among routes. In addition, the results also show that the spatiotemporal connectivity index used in our algorithm can reduce the computation time needed for searching for solutions without affecting the quality of the solutions.

This study evaluates the requisition–compensation balance of farmland (RCBF) policy with the telecoupling framework of "quantity-quality-productivity-ecological balance" in the middle reaches of Yangtze River Urban Agglomerations (MRYRUA) in China from 1995 to 2015. The RCBF index characterizes the quantity balance of farmland, while the quality balance is measured by examining the differences of physical elements (elevation, slope, soil texture, soil erosion, organic content, and pH value) between the requisitioned and compensatory farmland. The productivity balance is identified through the estimated grain yield, and the ecosystem services value (ESV) model is used to examine the ecological balance of the RCBF. The issues of "paddy field occupation and dry land compensation" and "superior occupation and inferior compensation" are proven to be critical in the MRYRUA, while the productivity of the requisitioned and compensatory farmland have reached a balance during the study periods. Two vicious circles and a serial of spillover effects are shown as the results of telecoupling. We suggest that more attention should be placed on land reclamation and land consolidation in order to avoid over-reliance on land exploitation. The findings have important implications for environmental protection, ecological compensation decisions, and land use sustainability.

INTRODUCTION In February 2014, the Bruton tyrosine kinase inhibitor ibrutinib received FDA approval for treatment of adult Chronic Lymphocytic Leukemia (CLL) patients (pts) treated with at least one prior therapy. Subsequently, ibrutinib was approved for treatment of CLL pts with 17p deletion (July 2014) and then as initial therapy for CLL (March 2017). Although ibrutinib has been shown to have durable responses, it is associated with bleeding complications and an increased risk of atrial fibrillation (Afib) in clinical trials. We present an analysis of Afib and stroke risk associated with ibrutinib in an observational study using real-world cancer registry data. METHODS: We used the SEER-Medicare database to identify adult Medicare beneficiaries (aged 66 years or older) diagnosed with CLL between 2008-2016, who were enrolled in Medicare at least 12 months prior to CLL diagnosis, and received their care through the traditional fee-for-service system (n=4,958). Sociodemographic characteristics, as well as date and cause of death were obtained from the SEER. Comorbidities listed on the Elixhauser comorbidity index were identified from inpatient and outpatient Medicare claims in the year prior to first treatment or first Ibrutinib use. Treatment modalities, including chlorambucil, rituximab, fludarabine, cyclophosphamide, bendamustine, and prescription of ibrutinib (starting in 2014), were identified from Medicare Part B and Part D data. We categorized primary exposure as: 1) pts who did not receive any treatment for CLL; 2) those who were treated but never given ibrutinib; and, 3) those treated with ibrutinib. Our primary outcomes were Afib and stroke. We compared demographic, clinical characteristics and comorbidities, and calculated incidence rates (IR, per 1,000 person-years) across the three treatment groups. Among those who were treated (n=2,789), we used Cox proportional hazards regression with inverse probability weighting (IPW) to estimate hazard ratios (HR) and 95% confidence intervals (95% CI) for the association between ibrutinib and other treatment with stroke and Afib, adjusted for age at treatment, sex, race, and Elixhauser comorbidity index. RESULTS: Among 4,958 CLL pts, 2,169 (44%) never received any treatment, 2,490 (50%) were treated but not with ibrutinib, and 299 (6%) received ibrutinib. Baseline characteristics are presented in Table 1. Importantly, pts treated with ibrutinib were younger at diagnosis (medianibrutinib, 75y; mediannon-ibrutinib, 76.5y; mediannot-treated, 80y) and at time of first treatment (medianibrutinib, 77y vs mediannon-ibrutinib, 78y). However, pts initially treated with ibrutinib (n=52) had higher age at first treatment, 77.5y (IQR: 72.5-82). Among pts who were never treated, the stroke incidence was 11.9% after CLL diagnosis (IR: 47.7/1,000 person-years) and Afib incidence was 15.2% (IR: 60.6/1,000 person-years). Among pts who received CLL treatment but not ibrutinib, the IR of stoke and Afib were 54.69/1,000 person-years and 67.2/1,000 person-years, respectively; and CLL pts treated with ibrutinib presented IR of stroke of 72.9/1,000 person-years while IR of Afib was 221.8/1,000 person-years. Among CLL pts who received treatment, after adjusting for potential confounding variables, the risk of Afib for pts treated with ibrutinib was 3-fold times the risk of pts whose treatment did not include ibrutinib (HR: 3.0; 95% CI: 2.1, 4.4). The association between ibrutinib versus other treatment with stroke was not statistically significant (HR: 1.1; 95% CI: 0.48, 2.3). CONCLUSIONS: Our analysis included data of CLL treatment patterns and therapy related toxicities from the initial years after ibrutinib regulatory approval. In alignment with clinical trial results, these early real - world data show ibrutinib pts have a statistically significant risk of Afib. Our analyses did not demonstrate an association between ibrutinib treatment and risk of stroke. We also found that pts prescribed therapy for CLL, including non - ibrutinib treatments, had a higher crude incidence of stroke and Afib than pts who had not initiated treatment. Future analyses will evaluate the risk of these complications in comparison with non CLL pts as well as their correlation with treatment in the setting of widespread use of BTK inhibitors and other targeted agents for treatment of CLL. Disclosures Koroukian: Cleveland Clinic Foundation: Research Funding; Celgene: Research Funding. Caimi:Celgene Corp: Other: Incyte Corporation - Ownership - Pharmacyclics, Inc. - Ownership - Celgene Corp. - Other, Speakers Bureau; Genentech: Research Funding; ADC Therapeutics: Research Funding.

This cross-sectional study examines temporal shifts in age distribution of obesity-associated cancers and non–obesity-associated cancers across race/ethnicity- and sex-specific strata., Key Points Question What are the temporal shifts in age distribution for obesity-associated cancers across sex- and race/ethnicity-specific strata? Findings In this cross-sectional study of 2 665 574 incident obesity-associated cancer cases and 3 448 126 incident non–obesity-associated cancer cases from 2000 to 2016, the percentage of individuals diagnosed with incident obesity-associated cancers increased in younger age groups, with some of the greatest increases observed for liver and thyroid cancers (all sex- and race/ethnicity-specific strata), gallbladder and other biliary cancers (non-Hispanic black men and women and Hispanic men), and uterine cancer (in Hispanic women in the 50- to 64-year age group). Meaning The findings suggest that there has been a shift of obesity-associated cancer burden to younger age groups and that interventions to reduce obesity and to implement individualized screening programs are needed., Importance Recent studies have documented an association between overweight and obesity and certain cancers, as well as an increased incidence of obesity-associated cancers (OACs) in younger individuals. However, although important for cancer-control efforts, it is not known which subgroups of the population are most affected by these changes. Objective To examine temporal shifts in age distribution of OACs and non-OACs across race/ethnicity- and sex-specific strata. Design, Setting, and Participants This population-based, cross-sectional study assessed individuals residing in Surveillance, Epidemiology, and End Results sites who were diagnosed with incident OACs and non-OACs from January 1, 2000, to December 31, 2016. Data analysis was performed from August 1, 2018, to June 30, 2019. Exposure Study years. Main Outcomes and Measures Changes in the age distribution of incident cases across race/ethnicity- and sex-specific strata over time. For all OACs and non-OACs, changes in the incidence rates, the number of incident cases, and the distribution of the cases across population subgroups were studied. Race/ethnicity- and sex-stratified logistic regression analysis was performed to determine whether the annual change in the odds associated with an age group increased (or decreased) to a greater (or lesser) extent in OACs than in non-OACs. Heat maps were created to highlight the change in the number of cases during the study period for each OAC and select non-OACs. Results The study population included 2 665 574 incident OAC cases (70.3% women) and 3 448 126 incident non-OAC cases (32.0% women). From 2000 to 2016, the distribution by age showed that the percentage of incident cases increased in the 50- to 64-year age group for both OACs and non-OACs. The increase in the number of OACs in this age group ranged from 25.3% in non-Hispanic white women to 197.8% in Hispanic men. The change in the number of OACs in the 20- to 49-year age group ranged from −5.9% in non-Hispanic white women to 94.6% in Hispanic women, and the increase in the number of OACs in the 65 years or older group ranged from 2.5% in non-Hispanic white women to 102.0% in Hispanic women. Logistic regression analysis revealed a greater annual increase in the odds for OACs than for non-OACs for individuals in the 50- to 64-year age group but a decrease for individuals in the 65 years or older group. Among men in the 50- to 64-year age group, the OAC to non-OAC ratio of odds ratios (ORs) ranged from 1.005 (95% CI, 1.002-1.008) in non-Hispanic black men to 1.013 (95% CI, 1.012-1.014) in non-Hispanic white men, implying that the annual increase was 0.5% higher for OACs than for non-OACs in non-Hispanic black men and 1.3% higher for OACs than for non-OACs in non-Hispanic white men. Among women in the 50- to 64-year age group, the OAC to non-OAC ratio of ORs ranged from 1.002 (95% CI, 0.999-1.006) in Hispanic women to 1.005 (95% CI, 1.002-1.009) in non-Hispanic black women. In men and women aged 65 or older, the OAC vs non-OAC ratio of ORs was consistently less than 1.000 for all race/ethnicity groups, indicating that whereas the OAC group experienced a decrease in this age group, the non-OACs experienced either a smaller decrease or an increase over time. Conclusions and Relevance This study indicates that from 2000 to 2016, a shift toward younger age groups occurred in incident OACs. The findings have important public health implications and suggest that interventions to reduce obesity and to implement individualized screening programs are needed.

1532 Background: In tandem with the Choosing Wisely initiative, ASCO’s Cost of Care Task Force has proposed a list of low-value (LV) procedures and therapies that may be of limited benefit to patients. Myelodysplastic syndrome (MDS) is the most common myeloid malignancy in the US. A complete diagnostic evaluation (CDE) of MDS requires a bone marrow biopsy, fluorescence in situ hybridization and chromosomal analysis. As a potential LV procedure, we evaluated receipt of CDE in MDS patients with isolated or no cytopenias and no transfusion dependence. Methods: Using national 2011-2014 Medicare data, we identified fee-for-service Medicare patients 66 years of age or older with an MDS diagnosis, one or no cytopenias, and no blood transfusions in the 16 weeks before or after an MDS diagnosis (n = 16,779). We examined the following variables that may have provided a clinical context to (or not to) pursue CDE – demographics (age, race, sex); number of Elixhauser comorbid conditions ( < 5 vs >5); nursing home status, prior history of lymphoma, myeloma, MGUS and other cancers; chronic kidney disease (CKD); colonoscopy; and therapies received including erythropoiesis stimulating agents (ESAs), hypomethylating agents (HMAs) or lenalidomide. We conducted Classification and Regression Tree (CART) analysis, a machine learning approach to identify combinations of factors in patients with little clinical justification for CDE, and Cox proportional hazards regression analysis to compare survival outcomes between those with or without CDE. Results: Over half of our study population (51%) received CDE. Of those, 46.6% were 80 years of age or older, 4.8% were nursing home residents; and 33.6% had 5 or more chronic conditions. Results from CART analysis showed that among patients with an isolated cytopenia (e.g., isolated anemia), 46.0% of patients >80 years (n = 860), and 57.7% (n = 1,156) of those in the 66-79 age group underwent CDE in the absence of CKD, colonoscopy, HMA, or ESA. Among those with no cytopenia (n = 3890), 866 patients received CDE in the absence of HMA, ESA, or history of lymphoma or progression to leukemia. In adjusted analyses, no survival benefit was associated with receipt of CDE (p = 0.24). Conclusions: A substantial number of patients with an MDS diagnosis, isolated or no cytopenias, and no transfusion dependence received a CDE in the absence of other diagnoses, procedures, or therapies that may have explained the clinical decision to perform CDE. These procedures entail costs, pain and anxiety, but do not appear to yield useful information to guide clinical management, as evidenced by the comparable survival outcomes between patients who did and did not undergo CDE. To promote patient-centered care, careful patient selection that reduces unnecessary CDE in MDS patients should be a priority in clinical decision-making.

1572 Background: Studies to date have shown post-Medicaid expansion (M-exp) decreases in the percentage of cancer patients who are uninsured and improvements in cancer stage at diagnosis in states that expanded Medicaid as part of the Affordable Care Act. However, most studies have examined impact of M-exp on stage outcomes at the population level, or among Medicaid and uninsured, rather than solely in the Medicaid population. Using cancer registry data from a non M-exp state (Georgia (GA)) and two M-exp states (Ohio (OH) and New Jersey (NJ)), we compared changes in cancer stage in patients on Medicaid, accounting for individual- and contextual-level characteristics at the Zip Code Tabulation Area (ZCTA) level. Methods: We used GA, OH, and NJ cancer registry data for individuals 20-64 years of age and diagnosed with incident invasive female breast (BC), cervical (CC), and colorectal cancer (CRC). Data spanned from 2010-2017 for GA and OH, and from 2011-2016 for NJ (for BC and CRC only), with 2014 marking the year in which Medicaid was expanded in OH and NJ. We retrieved demographic data (age, race/ethnicity, sex for CRC, insurance status, and cancer stage from the cancer registries), and obtained ZCTA-level data from the American Community Survey (e.g., income, education, and female-headed households). We defined late-stage diagnosis as regional- or distant- stage. We conducted multivariable logistic regression models by state and cancer site to examine changes in late-stage cancer diagnosis pre- and post-M-exp, accounting for individual- and ZCTA-level covariates. Results: The number of patients with incident cancer who were on Medicaid increased by 41.7% (n = 1757 to 2490), 59.6% (327 to 522), and 76.4% (953 to 1681) for BC, CC, and CRC cancers, respectively, in Ohio; by 92.4% (433 to 833) for BC and by over 100% for CRC (232 to 496) in NJ; but by 12.7% (662 to 746) among CRC patients in GA, where the number of BC and CC patients on Medicaid remained relatively stable. Adjusting for individual and contextual-level factors, the adjusted risk ratio (ARR and (95% Confidence Interval)) for late-stage disease was lowest for BC patients in OH (0.93 (0.87, 0.99)) and for CRC patients in GA (0.94 (0.89, 0.99)). The ARR for BC and CRC in NJ were not statistically significant, though they trended towards improvement. Similarly, changes in late-stage for CC were not statistically significant in OH or in GA. Conclusions: The increased number of cancer patients in Medicaid and the reductions in late-stage diagnosis observed may potentially translate into reduced, or at least stabilized, cancer-related morbidity and mortality burden among Medicaid beneficiaries over time. However, reductions in late-stage diagnosis were not consistent across cancer sites or states, possibly due to differences in population demographics, health behaviors, healthcare seeking patterns, and state-level cancer prevention efforts.

Study Objective: Genomic testing is essential to identify the best treatment modalities for certain cancers; yet, its uptake has been relatively low, especially among disadvantaged subgroups of the population. The aim of this study is to identify demographic and clinical factors associated with receipt of genomic testing in Ohioans diagnosed with either incident female breast, kidney, bladder, prostate, colorectal, or lung cancer. Methods: We used data from the 2009 linked Ohio Cancer Incidence Surveillance System and Medicare files, and identified genomic testing using the appropriate procedure codes in claims data. Our study population included 10,945 patients. Independent variables examined were age at diagnosis (< 65, 65-74, 75+), sex, race (White or All Other), dual enrollment in the Medicare and Medicaid program (or ‘dual’) as a marker for low income and heightened vulnerability, and advanced stage at diagnosis (Regional/Distant versus Local stage). We conducted multivariable logistic regression analysis to identify correlates of genomic testing by cancer site. Results: For all cancer sites combined, 11.1% were younger than 65, and 40.6% were older than 75 years of age. Eighty eight percent were White, 47.0% were women, 13.9% were duals, and one third were diagnosed with advanced-stage cancer. Overall, only 19.5% underwent genomic testing, ranging from a low of 6.7% in prostate cancer patients, to a high of 39.3% in breast cancer patients. In addition, we observed considerable variation in genomic testing by age, race, sex, dual status, and cancer stage across cancer sites. Adjusting for the independent variables listed above, being 75 years of age or older was significantly and positively associated with increased likelihood of undergoing genomic testing in breast (adjusted odds ratio: 1.17, 95% confidence interval: 1.04, 1.32), kidney and bladder combined (1.29 (1.09, 1.53)), and prostate cancer patients (1.45 (1.12, 1.89)). Advanced-stage disease was associated with increased likelihood of genomic testing in breast and colorectal cancer patients (1.40 (1.17, 1.67) and 3.07 (2.30, 4.11), respectively), but with decreased likelihood in kidney and bladder cancer patients (0.66 (0.48, 0.91)). Finally, we note that White patients with lung cancer were significantly more likely than others to undergo genomic testing (2.49 (1.29, 4.78)). Conclusion: Our data from 2009 provide baseline statistics on genomic testing uptake in Ohio. Data for subsequent years will help us to assess trends in providing personalized medicine. Citation Format: Siran Koroukian, Weichuan Dong, Johnie Rose, Fredrick Schumacher, Sarah Markt. Variations in genomic testing across cancer sites and by demographic characteristics [abstract]. In: Proceedings of the Twelfth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2019 Sep 20-23; San Francisco, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(6 Suppl_2):Abstract nr D074.

e17510 Background: The co-occurrence of cancer and end-stage-renal disease (ESRD) may pose significant challenge in the management of both diseases. Further complicating clinical decisions is our limited understanding of the comorbidity burden (CB), which also affects their outcomes to a considerable extent. The purpose of this study is to characterize the CB in prostate cancer patients, with and without ESRD. Methods: Using SEER-Medicare database between years 2000-2016, we retrieved sociodemographic variables, including age (40-54, 55-64, 65-74, and 75+), race (African American vs. all others), marital status (married/partnered vs. all others), residence in a census tract with poverty rate > 20%, and dual Medicare-Medicaid enrollment status; chronic conditions identified in the year of cancer diagnosis; and ESRD status preceding prostate cancer diagnosis. We limited our study population to cancer patients enrolled in Medicare at the time of cancer diagnosis, and were receiving their care through the fee-for-service system. In this descriptive analysis, we compared the prevalence of these conditions between prostate cancer patients by ESRD status. Results: Our study population included 2,046 ESRD and 302,136 non-ESRD men diagnosed with incident prostate cancer during the study period. Compared to non-ESRD patients, a disproportionately higher percentage of ESRD patients were in the 40-54 and 55-64 age groups compared to non-ESRD (11.0 vs 0.95% and 32.2% vs. 8.51%, respectively). Similarly, the percentage of prostate cancer patients who were African American was 44.1% among ESRD patients, compared with 13.6% in their non-ESRD counterparts. With regard to comorbidities, several conditions were significantly higher in ESRD than non-ESRD patients, including: anemia (65.4% vs. 15.3%), congestive heart failure (31.1% vs. 8.9%), ischemic heart disease (38.9% vs. 25.2%), diabetes (40.7% vs. 17.0%), hypertension (68.0% vs. 42.6%), hypothyroidism (4.6% vs. 2.9%), hyperlipidemia (43.1% vs. 35.1%), and stroke (3.7% vs. 2.5%). Conclusions: Compared to their non-ESRD counterparts, ESRD patients present with high CB, severely compromising their physiologic reserve and tolerance for various cancer treatment modalities, and affecting outcomes. Future studies should compare the prevalence of specific combinations of conditions constituting multimorbidity between ESRD and non-ESRD patients, and identify multimorbidity profiles associated with a lower likelihood to receive standard treatment. Such detailed analysis will be foundational to clinical management and outcome studies.

e17513 Background: Persons with end-stage renal disease (ESRD) comprise a vulnerable subgroup of the population. Yet, rarely have studies on cancer disparities accounted for ESRD status. We sought to compare prostate cancer stage at diagnosis, treatment, and outcomes among men with and without ESRD. Methods: Utilizing SEER-Medicare database between years 2000-2016, we retrieved sociodemographic variables, including age (40-54, 55-64, 65-74, and 75+), race (African American vs. all others), marital status (married/partnered vs. all others), residence in a census tract with poverty rate > 20%; cancer stage at diagnosis (local, regional, distant, and unknown); treatment (surgery and radiation therapy); and overall and cancer specific survival (OS and CSS) from the SEER, and ESRD status preceding prostate cancer diagnosis from the Medicare portion of the file. In addition to descriptive analysis, multivariable survival analysis was utilized to evaluate the association of ESRD status with OS and CSS after adjusting for patient covariates. Results: Our study population included 2,922 ESRD and 600,848 non-ESRD men diagnosed with incident prostate cancer during the study period. A disproportionately higher percentage of ESRD patients were in the 40-54 age group compared to non-ESRD (9.7 vs 3.6%). Similarly, the percentage of prostate cancer patients who were African American was 41.5% among ESRD patients, compared with 14.4% in their non-ESRD counterparts. In addition, more ESRD than non-ESRD patients lived in census tracts with poverty rates greater than 20% (31.2% vs. 18.0%). With regard to stage, 10.6% of ESRD patients but only 4.5% of non-ESRD patients were diagnosed with distant-stage disease; and while only 24.8% of ESRD patients underwent surgery, 36.9% of non-ESRD patients did so. The rate of radiation therapy was similar across the two groups. Multivariable survival models showed that, adjusting for patient covariates, ESRD status was associated with unfavorable OS and CSS (adjusted hazard ratio (AHR): 4.00 (95% confidence interval: 3.82, 4.18), and AHR: 1.54 (1.37, 1.73), respectively). Median follow up was 3.35 and 7.13 years between the ESRD and non-ESRD cohorts. Conclusions: ESRD status in prostate cancer patients is associated with marked sociodemographic and economic disparities. Compared to their non-ESRD counterparts, more ESRD patients are diagnosed with advanced-stage disease, are less likely to undergo surgery, and experience shorter overall and cancer-specific survival. Future studies should be directed to gain insight into factors contributing to unfavorable cancer-specific outcomes.

7555 Background: A complete diagnostic evaluation (CDE), including bone marrow (BM) biopsy, cytogenic testing, FISH panel, and/or flow cytometry, is widely accepted as a prerequisite for histopathologic confirmation of MDS. Earlier reports have raised concerns regarding accuracy of MDS diagnosis in population registries that do not require confirmatory tests for disease reporting. We queried Medicare files to analyze the extent of use of, and factors associated with CDE for diagnosing MDS. Methods: The study population included Medicare patients with at least one inpatient or two outpatient claims for MDS within a 12-month period during the years 2012 and 2013. Variables included age, sex, race, morphologic MDS categories, cytopenias (isolated or any combination), transfusion burden, comorbididities, as well as county-level characteristics (income, educational attainment, rurality, and availability of internal medicine subspecialists). Classification and regression tree (CART) and multivariable logistic regression analysis were used to identify combinations of factors associated with receipt of CDE. Results: Our study population included 45,067 MDS patients, of whom only 68.6% received CDE that included BM biopsy and/or chromosomal studies. The percent of patients undergoing CDE was significantly lower among those 85 years of age or older (56.1%); women (63.7%); Blacks (64.2%); patients with isolated anemia (60.3%); and those who were transfusion independent (p < 0.001 for all comparisons). On the other hand, CDE was significantly higher among patients with the category of MDS with excess blasts (89.4%); those presenting with pancytopenia (92.2%); and those with high transfusion burden (80%). There was little variation by comorbidity burden, county level characteristics, or availability of subspecialists. In CART analysis, 80% of patients presenting with any two or more cytopenias received CDE. In multivariable analysis, advanced age, female sex, Black race, higher comorbidity burden and availability of subspecialists were associated with a lower likelihood to receive CDE, while patients with MDS with excess blasts, and with any > 2 cytopenias were more likely to receive CDE ( P< .0001). Conclusions: Nearly one third (31%) of MDS patients in the Medicare database lack CDE especially BM biopsy. Disparities in CDE can be corrected through proper physician education and stringent registry reporting to avoid such high diagnostic inaccuracies, leading to potential missed treatment opportunities.