Abstract
Introduction:
Previous studies have shown that uninsured and Medicaid patients had higher morbidity and mortality due to limited access to healthcare. Disparities in cancer-related treatment and survival outcome by different insurance have been well established (Celie et al. J Surg Oncol.,2017). There are approximately 8,260 newly diagnosed HL cases in the US yearly (Master et al. Anticancer Res.2017). Therefore, we aim to investigate the variation of survival outcome and insurance status among HL patients.
Methods:
We extracted data from the US National Cancer Institute's Surveillance, Epidemiology and End Results (SEER) 18 program. HL patients who were diagnosed from 2007-2014 were included. Demographic information including age, sex, race, annual household income, education and insurance were also collected. Insurance includes uninsured, insured and any Medicaid. Race/ethnicity includes white, black and other (including American Indian/AK native, Asian/Pacific Islander). HL is categorized by using International Classification of Disease for Oncology (ICD-O-3) into classical HL NOS (CHL NOS), nodular lymphocyte predominant HL (NLP), lymphocyte rich (LR), mixed cellularity (MC), lymphocyte depleted (LD), and nodular sclerosis (NS). Treatment modality included RT alone, CT alone, RT and CT combined, and no RT or CT. Survival time was estimated by using the date of diagnosis and one of the following dates: date of death, date last known to be alive or date of the study cutoff (December 31, 2014). Chi-square test and multivariate Cox regression were performed by using SAS 9.4 (SAS Institute Inc., Cary, NC, USA).
Exclusion criteria include: 1) patients with unknown or unspecified race; 2) patients who survived less than 6 months because time of radiotherapy/chemotherapy was not known to the time of diagnosis; 3) patients with any other type of cancer prior to the diagnosis of HL; 4) patients with second or later primaries, and who were not actively followed.
Results:
A total of 14.286 HL patients were included in the analysis. Table 1 indicates the insurance status and demographic and tumor characteristics among HL patients diagnosed between 2007 and 2014. Patients with black race, male sex, and B symptoms were more likely to be uninsured and on any Medicaid compared to other races, female sex and without B symptoms (p<0.01). As stage of disease increased, the percentage of insured patients decreased from 82.0% to 71.7%, (p<0.01). As with year of diagnosis advanced, the percentage of uninsured did not appear to be changed however the proportion of both those with insurance and any Medicaid decreased slightly by 2.4% (p<0.01). Those who received RT only were most likely to have insurance (89.6%) followed by combination modality (80.1%). As expected, uninsured status was associated with lower income and education level (p<0.01).
Table 2 shows the insurance and hazard ratio among HL patients by year of diagnosis adjusting for race, sex, histology type, income, education, and year of diagnosis. Any Medicaid patients had the highest HR of death from 2007-2010 compared to insured patients. Without insurance was also associated with increased risk of death but only significant in 2008, HR=2.26, 95% CI (1.35, 3.80).
The survival outcomes comparing different insurance status by age groups (<=29 and 30-64) were demonstrated in Kaplan-Meier Curve. In the age 29 or less group, insured patient showed has the best survival outcome followed by any Medicaid and then the uninsured. In the age 30-64 group, Medicaid patients had the worst survival outcome compared to those with or without insurance.
Conclusion:
Insurance status is one of the most important contributors of health disparity, especially in malignancy given the significant financial toxicity of therapies. We found that the proportion of the uninsured was trending up before the Affordable Care Act (ACA). Regarding the HL outcome, insured patients had the best survival across all age groups even though not significantly while Medicaid patients had the worst outcomes in almost all age groups, even worse than the uninsured after adjusting for the disease stage at diagnosis and sociodemographic factors. It would be of interest to explore the reason behind Medicaid patients' relatively poor outcomes. Future studies may also investigate how ACA, Medicaid expansion, and the possible upcoming republican healthcare reform influence HL outcome.
Disclosures
No relevant conflicts of interest to declare.
Concordance of Disclosed Hospital Prices With Total Reimbursements for Hospital-Based Care Among Commercially Insured Patients in the US
