Disparities in health care determine prognosis in newly diagnosed glioblastoma.

OBJECTIVE Glioblastoma (GBM) is an aggressive brain malignancy with a short overall patient survival, yet there remains significant heterogeneity in outcomes. Although access to health care has previously been linked to impact on prognosis in several malignancies, this question remains incompletely answered in GBM. METHODS This study was a retrospective analysis of 354 newly diagnosed patients with GBM who underwent first resection at the authors' institution (2007-2015). RESULTS Of the 354 patients (median age 61 years, and 37.6% were females), 32 (9.0%) had no insurance, whereas 322 (91.0%) had insurance, of whom 131 (40.7%) had Medicare, 45 (14%) had Medicaid, and 146 (45.3%) had private insurance. On average, insured patients survived almost 2-fold longer (p < 0.0001) than those who were uninsured, whereas differences between specific insurance types did not influence survival. The adjusted hazard ratio (HR) for death was higher in uninsured patients (HR 2.27 [95% CI 1.49-3.33], p = 0.0003). Age, mean household income, tumor size at diagnosis, and extent of resection did not differ between insured and uninsured patients, but there was a disparity in primary care physician (PCP) status-none of the uninsured patients had PCPs, whereas 72% of insured patients had PCPs. Postoperative adjuvant treatment rates with temozolomide (TMZ) and radiation therapy (XRT) were significantly less in uninsured (TMZ in 56.3%, XRT in 56.3%) than in insured (TMZ in 75.2%, XRT in 79.2%; p = 0.02 and p = 0.003) patients. Insured patients receiving both agents had better prognosis than uninsured patients receiving the same treatment (9.1 vs 16.34 months; p = 0.025), suggesting that the survival effect in insured patients could only partly be explained by higher treatment rates. Moreover, having a PCP increased survival among the insured cohort (10.7 vs 16.1 months, HR 1.65 [95% CI 1.27-2.15]; p = 0.0001), which could be explained by significant differences in tumor diameter at initial diagnosis between patients with and without PCPs (4.3 vs 4.8 cm, p = 0.003), and a higher rate of clinical trial enrollment, suggesting a critical role of PCPs for a timelier diagnosis of GBM and proactive cancer care management. CONCLUSIONS Access to health care is a strong determinant of prognosis in newly diagnosed patients with GBM. Any type of insurance coverage and having a PCP improved prognosis in this patient cohort. Higher rates of treatment with TMZ plus XRT, clinical trial enrollment, fewer comorbidities, and early diagnosis may explain survival disparities. Lack of health insurance or a PCP are major challenges within the health care system, which, if improved upon, could favorably impact the prognosis of patients with GBM.