Search

10574 Background: There is growing evidence that adult oncology patients who know they are dying choose less intense care. Further, high intensity care is associated with worse caregiver outcomes. Little is known about pediatric oncology end-of-life care intensity. Methods: Using the California Office of Statewide Health Planning and Development administrative database linked to death certificates, we performed a retrospective population based analysis of cancer patients aged 0-21 who died between 2000 and 2011. The frequency of previously defined end-of-life intensity markers (hospital death, intense medical interventions, IV chemotherapy, and gastrostomy and tracheostomy tube placement) were calculated and multivariable logistic regression was used to determine clinical and sociodemographic factors associated with > 2 intensity markers (as above), intense medical intervention (cardiopulmonary resuscitation, intubation, ICU admission, or hemodialysis), and hospital death. Results: The 3,732 pediatric cancer decedents were 34% non-Hispanic whites and 45% Hispanic; 41% had hematologic malignancies and 59% solid tumors. The most prevalent intensity markers included: hospital death (63%) and ICU admission (20%). 65% had > 1 intensity marker, 23% > 2, and 22% > 1 intense medical intervention. There was a bimodal association between age and intensity: the youngest patients (age < 5) and adolescent patients (age 15-21) were more likely to receive intense care: < 5y (intense medical intervention: OR = 1.42; 95% CI, 1.1-1.9; hospital death: OR = 1.72; 95% CI, 1.4-2.2; > 2 markers: OR = 1.37, 95% CI 1.1-1.8); 15-21y (intense medical: OR = 1.48; 95% CI, 1.2-1.9; hospital death: OR = 1.39; 95% CI, 1.1-1.7; > 2 markers: OR = 1.35, 1.1-1.7) (references: 5-9y). Other factors associated with intensity included, hematologic malignancies, minority status, and death between 2008 and 2011 vs. < 2008. Conclusions: Nearly two-thirds of the pediatric cancer decedents had ≥1 marker of intense care and disparities exist. Patients < 5 and adolescents were more likely to receive intense end-of-life care. Further research needs to determine if these rates and variation are consistent with patient goals and factors associated with goal concurrent care.

6561 Background: Private health insurance is associated with improved outcomes in cancer patients. We know little, however, about the impact of the ACA-DCE, which extended private insurance to young adults (up to age 26) beginning in 2010, on the insurance status of young adults with cancer. This study sought to determine the effect of the ACA-DCE on having private insurance coverage among hospitalized young adult oncology patients. Methods: We performed a retrospective, population-based analysis of hospitalized young adult oncology patients (22-30 years-old) in California during 2006-2014 (n = 11,062) using the Office of Statewide Health Planning and Development database. Multivariable regression analyses examined the social and clinical predictors of having private insurance. Results are presented as adjusted odds ratios (OR) and 95% confidence intervals (CIs). A difference-in-difference analysis examined the influence of the ACA-DCE on insurance coverage by race/ethnicity and zip code federal poverty level. Results: Multivariable regression demonstrated patients of black and Hispanic race/ethnicity were less likely to have private insurance both before and after the ACA-DCE, compared to non-Hispanic white patients. Younger age (22-25 years) was associated with having private insurance after the ACA-DCE implementation (OR 1.18, CI 1.05-1.33; reference, 27-30 years). In the difference-in-difference analysis, private insurance increased among non-Hispanic whites aged 22-25 living in medium- (2006-2009: 64.6% versus (vs) 2011-2014: 69.1%; p = 0.003) and high-income zip codes (80.4% vs 82%; p = 0.043) and among Asian patients aged 22-25 living in high-income zip codes (73.2 vs 85.7%; p = 0.022). Private insurance decreased for all Hispanic patients aged 22-25 between the two time periods. Conclusions: The ACA-DCE provision was an important first step in increasing coverage, but it was not universal and generated disparity in coverage as gains occurred for non-Hispanic white and Asian patients living in higher income zip codes. This policy change was shown to increase coverage for a traditionally underinsured population and attention should now focus on those remaining uninsured.

132 Background: Cancer is the leading cause of non-accidental death amongst adolescents and young adults (AYA), aged 15-39, in the U.S. It is critical to understand end of life (EOL) care of AYA cancer decedents, including use of medically intense interventions like intubation. Although desired by some, most patients prefer a natural death. We sought to determine rates of medically intense interventions at end of life for AYA cancer decedents and associated factors. Methods: Using the California Office of Statewide Health Planning and Development private administrative database linked to death certificates, we performed a retrospective population based analysis of patients aged 15-39 with cancer who died between 2000-2010. We used previously defined administrative codes indicative of intense EOL care: intubation, CPR, hospital re-admission, and ICU admission in the last 30 days of life, and location of death. The frequencies of each intense item were calculated and multivariate logistic regression was used to determine clinical (including treatment at specialty center vs non) and socio-demographic factors associated with each item and receipt of ≥ 2 items. Results: The 8,978 AYA cancer decedents were 46% non-Hispanic white, 29% Hispanic, 10% non-Hispanic black, 11% Asian; 21% had hematologic malignancies, 70% had solid tumors, and 9% had secondary neoplasms; 58% were hospitalized only at non-specialty centers in the last 6 months of life. 62% received ≥ 1 medically intense EOL care intervention, and 32% received > 2. Factors associated with > 2 intense EOL care interventions were: non-Hispanic black (OR 1.38, 95% CI 1.16-1.65), Hispanics (1.19, 1.06-1.35), Asians (1.30, 1.10-1.52); those sometimes (2.19, 1.87-2.56) or never (1.44, 1.26-1.65) seen at specialty centers; hematologic malignancies (1.77, 1.57-2.00 ref grp: solid tumors) whereas secondary malignancies were not associated with > 2 intense markers (0.68, 0.56-0.83). Conclusions: Nearly two-thirds of the AYA cancer decedents received medically intense EOL interventions and disparities exist in receipt of such care. Further research needs to determine if the disparities are due to healthcare system, patient preference, or other factors.

126 Background: Cancer is the leading cause of non-accidental death among adolescents and young adults (AYA) in the US. The last year of life involves significant inpatient resources for older patients. We sought to determine the quantity and pattern of inpatient care for AYA cancer decedents in the last year of life to learn their healthcare system impact and groups to target for intervention. Methods: Using the California Office of Statewide Health Planning and Development private administrative database linked to death certificates, we performed a retrospective cross-sectional population-based analysis of patients aged 15-39 with cancer who died between 2000-2011. We calculated the number of admissions, hospital days, and percent of the cohort admitted each day in the last year of life. We determined the bed-day distribution across the population and the clinical and socio-demographic factors associated with high inpatient utilization. Results: The 9562 AYA cancer decedents were 45% non-Hispanic white, 30% Hispanic; 20% had hematologic malignancies, 70% had solid tumors. They were hospitalized, on average, 4.3 times and 40.6 days in the last year of life, increasing 3 months before death. Bed day occupation was skewed: 5% of the patients occupied 20% of the bed days and 18% occupied 50%. Factors associated with increased odds of being a top 5% utilizer were being Hispanic (OR: 1.48, 95% CI: 1.15-1.90, ref: non-Hispanic white), and having Acute Myelogenous Leukemia (AML) (2.7, 1.74-0.67, ref: Acute Lymphoblastic Leukemia (ALL). Factors associated with decreased odds were rural residence (0.39, 0.23-0.67, ref: urban), HMO insurance (0.58, 0.38-0.88, ref: private non-HMO), no hospitalizations at specialty centers (0.72, 0.53-0.97, ref: specialty center only), and having lymphoma or a solid tumor (ORs 0.035-0.39, CI range 0.005-0.56, ref: ALL). Conclusions: AYAs dying of cancer spent 40.6 days in the hospital in their last year of life (more than 1 in 9 days) with higher rates for Hispanics, AML patients, those with urban residence, and those hospitalized at specialty centers. Further research needs to determine what is driving these patterns, if they are preventable, and if they align with patient and family wishes.

e18020Background: AYAs with cancer have had less improvement in survival compared to their younger counterparts. This may be partly due to differential access to specialized cancer centers (SCC), a...

e17504 Background: Previous analysis has shown that Hispanic pediatric oncology patients in California have decreased utilization of pediatric cancer specialty centers (PCSC). The purpose of this study was to examine the clinical and non-clinical factors associated with decreased inpatient utilization of PCSCs by Hispanic pediatric oncology patients. Methods: Using the California Office of Statewide Health Planning and Development private database we performed a retrospective population based analysis of Hispanic patient discharges 0-18 years of age to explore the influence of age, gender, payer, income, distance to admission site, tumor type and diagnosis of fever and neutropenia on the utilization of PCSCs by Hispanic patients. Results: Analysis of 82,832 Hispanic pediatric oncology discharges showed concentration of hospitalization in PCSCs increasing from 64% in 1983 to 94% in 2010. In addition, an increased percentage of acute lymphoblastic leukemia discharges was seen for Hispanics (28.3%) compared to all discharges (22.4%). The adjusted regression analysis revealed decreased utilization of PCSCs for patients ages 15-18 (OR 0.466; CI 0.441-0.493) and for those residing 6 - 10 miles from a PCSC (OR 0.927; CI 0.883 - 0.974). Increased utilization of PCSCs was seen for Hispanic patients living in zip codes where household income was >4 times the Federal Poverty Level. Payer status was not significantly associated with discharge from a PCSC. None of these findings changed when Los Angeles County was removed from the analysis. A separate analysis was done on Hispanic pediatric oncology patients with a discharge diagnosis of fever and neutropenia. For this subset decreased utilization was seen for those with private insurance (OR 0.84; CI 0.81-0.887), age 40 miles from a PCSC (OR 0.913; CI 0.854-0.975). Conclusions: Utilization of inpatient pediatric oncology care by Hispanic patients in California has become regionalized in PCSCs since 1983 and is influenced by age, and geography, with payer status not being a significant contributor. Clinical outcomes as they relate to access to PCSCs for Hispanic patients should be examined.