Utilization of Palliative Care Among Patients with Hematologic Malignancies: An Extensive National Analysis to Define Trends and Patterns

Background: There is growing recognition of palliative care (PAL) as an integral aspect of cancer treatment. Several guidelines recommend universal access and early utilization of PAL as a goal for patient care but no data exists defining current trends and practice patterns for PAL use, especially in hematologic malignancies.

Methods: The 2016 National Cancer Database (NCDB) was utilized for patients with non-Hodgkin lymphoma (NHL), multiple myeloma (MM), chronic lymphocytic leukemia (CLL), acute myeloid leukemia (AML) and Hodgkin disease (HD) diagnosed between 2004-2013. Deceased patients who received PAL as part of initial therapy for these diagnoses were stratified by survival duration (<6, 6-24, 24-60 and 60+ months). These strata were then utilized to determine associations between PAL usage and socio-demographic, disease and facility characteristics using multivariate logistic regression analysis.

Results: A total of 293518 patients were present in NCDB for the identified diagnoses (AML: 55827, CLL: 20634, HD: 9871, NHL: 147614, MM: 59572). Of these 15352 (5.2%) received PAL as part of initial therapy. Compared to patients diagnosed in 2004-2005, patients diagnosed in 2012-2013 had significantly higher odds of PAL use in those surviving <6 months (OR 1.28, 95% CI 1.17, 1.39), while not for those with longer survival. Type of PAL used included surgery/radiation/chemotherapy (S/R/C) in 68%, pain management only in 14.5%, both S/R/C/ and pain management in 5% and unknown modality in 12.5%. There was a significant decrease in overall PAL use as duration of patient survival increased, from 7.2% in <6 month to 2.9% in 60+ month group (p<.001). Among the selected diagnoses, PAL use was highest for MM patients for all survival groups (OR 2.65-59.56, ref=AML) and lowest for CLL (OR 0.51-2.94, ref=AML). PAL use increased significantly for the <6 month and 6-24 month survival groups with increasing age (p<.001) but not for those with longer follow up. There was no difference in PAL use by gender for any survival group. PAL utilization increased significantly for patients with higher Charlson-comorbidity index but only for those with <6 month survival (p<.001), not for those with longer survival. PAL use was significantly different by race-ethnicity in those who survived <6 month with lesser use in non-Hispanic Blacks (OR 0.90, 95% CI 0.82, 0.98) than non-Hispanic Whites while not different in those with longer survival. Within a racial-ethnic group, Hispanics had an increased utilization of PAL for those with improving survival (60+ month group OR 1.7, 95% CI 1.09, 2.63). There was no difference in PAL use by urban/rural location of patient residence or distance from the treating facility for any survival duration group. The only influence of geographic region was that for patients with <6 month survival, PAL utilization was higher in Mountain (OR 1.44, 95% CI 1.27, 1.64) and Pacific (OR 1.19, 95% CI 1.09, 1.32) regions (ref=East Coast). Socioeconomic factors affecting PAL use included insurance status but only for those surviving 6-24 months where patients with "other government" payer were most likely to get PAL (OR 1.82, 95% CI 1.32, 2.5, ref=private insurance). There was a significant interaction of median patient income and PAL use across survival groups (p=0.011) but no clear trend within any specific subgroup by income or survival category. Higher PAL use was noted with increasing literacy level (p<.001) only in those with <6 month survival and not for any other survival group. Compared to non-academic cancer programs, academic/research programs were less likely to utilize PAL in those surviving 24-60 months (OR 0.85, 95% CI 0.76, 0.94) and 60+ months (OR 0.74, 95% CI 0.61, 0.89), while there was no significant difference in those surviving <6 months (OR 1.06, 95 % CI 1, 1.12) or 6-24 months (OR 1, 95% CI 0.92, 1.08).

Conclusions: We present the largest analysis so far of utilization of PAL in patients with hematologic malignancies. Despite national guidelines and near universal recommendations, overall utilization of PAL was dismal and lesser so in academic cancer programs including NCI-designated centers. There was no clear subgroup showing uniform utilization, suggesting sporadic use at best. We noted significant heterogeneity in practice patterns by all characteristics studied underscoring the need for standardized implementation with public, healthcare provider, institutional and political will.