The study was exempt from Institutional Review Board review because the dataset is publicly available and de-identified. The SEER program was utilized to examine the effects of insurance status on CVD mortality in a non-elderly cohort in our study. SEER is a registry representing 28% of the US population based on yearly tumor census data, including patient demographics, primary tumor site, stage at diagnosis, initial course of treatment, follow-up time, and survival [11]. The SEER registry has also collected data on patient’s insurance status since 2007, but is not validated for patients aged more than 65 years old due to Medicare eligibility.
Participants were included if they met all of the following criteria 1) aged 18 to 64 years at the time of diagnosis of an initial malignancy; 2) having the following cancers (colorectal, lung, breast, ovarian, cervical, bladder cancer, Hodgkin’s lymphoma and Non-Hodgkin’s lymphoma); 3) diagnosis of cancer after 2007. The exclusion criteria included 1) lack of insurance information, 2) lack of information on cause of death.
Insurance status was defined as insured or insured without specifics, Medicaid Insurance (any Medicaid, including Indian Health Service), or uninsured. The definition for insured in SEER included private insurance, Medicare, or military coverage at the time of diagnosis and will be referred as Private Insurance going forward for simplification purposes. Race was classified as Hispanic and non-Hispanic White, non-Hispanic Black, Asian and Others. We categorized patients into three groups by the year of diagnosis (2007- 2010, 2011-2013, 2014-2016) to assess changes in insurance status prior to and after implementing the Affordable Care Act (ACA). The extent of disease was categorized as in situ, local (no nodal or metastatic disease), region (nodal disease), or distant (any metastatic disease). The initial course of treatment was categorized based on whether patients received chemotherapy, radiotherapy or surgery as binary variables for each treatment.
Cause of death was defined by the International Classification of Diseases (ICD-9) code based on death certificates. CVDs was defined as heart disease, hypertension, cerebrovascular disease, atherosclerosis, aortic aneurysm/dissection, and other diseases of arteries, arterioles, or capillaries.
Patient and clinical characteristics were compared using the Chi-Square test. A competing risk model was conducted and communicative incidence function was calculated with the Fine and Grey model. We also used the Fine-Grey model to compare the risk of death from CVD among different insurance groups, after adjusting other risk factors including age, gender, marital status, race/ethnicity, year of diagnosis, if patient received radiation therapy, chemotherapy or surgery respectively, and cancer sites. The proportional hazards assumption was confirmed by inspection of log (-log [CVD death] curves).
To further illustrate the risk of death from CVD in cancer patients in different insurance group, we used standardized mortality ratios (SMRs) to provide the relative risk of death from CVD for cancer survivors as compared to the standard US population, adjusted by age, race, and sex over the same time. We describe the risk of death from CVD as a function of age at cancer diagnosis, year of cancer diagnosis, and follow up time, respectively. We also describe the risk of CVD mortalities by cancer site. The reference cohort was US mortality as reported in the National Vital Statistics System and maintained by the National Center for Health statistics. The SMRs were calculated using SEER*Stat 8.3.6. and all other statistical analyses were performed using SAS 9.4 software and Microsoft Excel 16.0.