ABSTRACT
Purpose: In the state of Wisconsin, breast cancer patients from African American (AAs) communities have lower survival rates compared to their Caucasian counterparts. Multiple inequities related to sociodemographic factors, delays in diagnosis, advanced disease stage at presentation and presence of comorbidities including higher body mass index (BMI) contribute to these disparities, many of which have only widened during the COVID-19 pandemic. This study examined specific factors related to prolonged hospital length of stay (LOS) for breast cancer patients admitted to inpatient units during the pandemic. Methods: This analysis includes initial CY20 LOS medical record data for hospitalized patients 18 years and older with a diagnosis of breast cancer from 1/1/2020-12/31/2020. Supplemental data included disease registry and diagnostic data, and SES data determined by patient zip code. Poisson regression models with robust standard errors were used to compare the LOS index (LOSi) between groups of patients based on race, SES group, primary payer, and BMI. Results: A total of 272 patients with breast cancer that were admitted to inpatient oncology units were identified. Demographics included White (72.4%), Black (22.4%), and others (5.1%). Other characteristics included: low SES (8.8%), medium-low (9.5%), medium (15.4%), medium-high (11.0%), high SES (4.4%), and others (non-SMilwaukee county) (50.7%), Medicaid (8.8%), Medicare (61.3%), Managed care (29.0%), and others (0.73%). Body mass varied among the patients;underweight (0.36%), overweight (30.8%), obese (41.5%). There were significant differences in LOSi: Black (LOSi=1.24, p=0.01), medium-low SES (LOSi=1.46, p=0.02), Medicaid (LOSi=1.40, p=0.00), underweight (LOSi=1.66, p=0.00), and overweight (LOSi=1.23, p = 0.01) patients had slightly longer LOSi, with LOSi ratio above 1. Conclusion: This study shows how patient-specific factors such as race, SES, primary payer, and BMI contribute to inpatient LOS for breast cancer patients. Healthcare systems may benefit by addressing indicators and patients' factors to reduce hospital LOS, and ultimately healthcare costs.
ABSTRACT
Background: Patients with cancer have increased risk of developing SARS-Cov-2 (COVID-19) infection. It is unknown if characteristics related to breast cancer increase the risk of COVID-19 infection. In this retrospective matched case control study, we aim to identify breast cancer related risk factors associated with developing COVID-19 and describe outcomes of patients with breast cancer diagnosed with COVID-19. Methods: Women with breast cancer treated at the Medical College of Wisconsin and diagnosed with COVID-19 between March and December 2020 served as cases. Women with breast cancer without COVID-19 diagnosis within the same time frame were identified as potential controls. Controls were chosen by matching for age (≥60 vs <60), obesity (BMI <30 vs ≥30), county (Milwaukee vs suburban), race (white vs non-white) and diabetes mellitus (DM) with 3:1 matching planned. Univariate comparisons between cases and controls were done via RaoScott stratified chi-square test for categorical outcomes and stratified t-test for continuous variables. Conditional logistic regression was done to evaluate the joint effect of multiple characteristics on the odds of being a COVID-19 case. Results: Twenty-five cases and 77 controls were identified. All cases were fully matched by age, obesity, county, and race with 3 cases not able to be matched for DM. Mean age was 54.6 vs 54.9 (p=0.88), BMI 31.0 vs 31.6 (p=0.69), 48% lived in Milwaukee county and 68% were white (cases 24% black & 8% American Indian;controls 32% black). Regarding COVID outcomes, 24.0% (n=6) of cases were hospitalized, median length of stay was 2 days, 8% (n=2) needed oxygen, 4% (n=1) were intubated and 4% (n=1) died due to COVID-19. COVID-19 led to treatment delays in 40% of cases. On univariate analysis of cases vs controls, 64 vs 75% were ER/PR+ (p=0.31), 6.5 vs 5.2% HER2+ (p=0.34), and 9.0 vs 4.2% triple negative (p=0.10). There were no significant differences in breast cancer stage. At time of COVID diagnosis (or last clinic contact if control), 16 vs 14% had active disease (p=0.81), 72 vs 74% were on active treatment (p=0.85), with 21 vs 4% being on chemotherapy (p=0.007), and 44 vs 52% on endocrine therapy (p=0.49). On conditional logistic regression, being on active chemotherapy (OR 5.8, p=0.043) significantly increased the likelihood of developing COVID with a trend seen for triple negative disease (OR 2.8, p=0.12). Conclusions: In this matched case control study of patients with breast cancer, active chemotherapy was significantly associated with an increased likelihood of developing COVID-19 with a trend seen for triple negative disease. Rates of death due to COVID-19 were overall low. Our analysis was limited by small numbers and an inability to fully match patients for DM. These findings support continued strict precautions for those on active chemotherapy and warrants further analysis in those with triple negative disease.