Among patients with stage IV disease, there was a significant increase in the utilization of palliative care, from 12.9% in 2003 to 19.2% in 2011 (P<0.001). Considering the various interventions individually, the rate of surgery significantly decreased over this time interval, while the rate of radiation therapy utilization remained stable (Figure 1). Systemic therapy, pain management strategies, and combination therapy all significantly increased over this time interval (Figure 1). Referral for palliative care also increased, although the type of therapy that was performed, if at all, was unable to be determined among this subgroup.


(To view a larger version of Figure 1, click here.)

Logistic regression

After controlling for baseline differences between the cohorts, multivariable logistic regression showed that increasing age was associated with lower odds of receiving palliative therapy, particularly among those over 80 years (Table 2). The difference in utilization between males and females was not statistically significant (P=0.335). Being of black race or Hispanic ethnicity, however, was associated with lower utilization of palliative measures [odds ratio (OR) 0.83 with 95% confidence interval (CI) 0.77 to 0.89, and OR 0.80 with 95% CI: 0.75 to 0.85, respectively] versus Caucasians. Presence of associated comorbidities increased the likelihood of receiving palliative therapy, with 16% higher odds in those with one comorbidity (95% CI: 1.10 to 1.21), and 27% higher odds in those with two or more comorbidities (95% CI: 1.18 to 1.37). Compared to patients with government insurance, utilization was lower for privately insured patients (OR 0.90; 95% CI: 0.85 to 0.96). Comprehensive community and academic centers were more likely to offer palliative therapies than community cancer centers (Table 2). Additionally, significant regional variation was noted, with the Northeast region of the country having the highest utilization compared to the South, Midwest, or West (Table 2).


(To view a larger version of Table 2, click here.)

Survival analyses

Kaplan-Meier curves for the various types of palliative intervention are shown in Figure 2. After controlling for patient- and hospital-related factors using a multivariable Weibull model, palliative radiation did not significantly affect survival of patients with stage IV disease (Table 3). Both surgery [hazard ratio (HR) 0.94; 95% CI: 0.90 to 0.98] and systemic chemotherapy (HR 0.55; 95% CI: 0.53 to 0.57) were associated with improved survival. Pain management alone, combination therapy, and referrals without a designated intervention were all associated with worsened survival, when compared to patients who received neither palliative nor non-palliative interventions (Table 3). All forms of non-palliative interventions (surgery, radiation, chemotherapy, and combination therapy) were associated with improved survival (Table 3).


(To view a larger version of Figure 2, click here.) 


(To view a larger version of Table 3, click here.)