Initial treatment is associated with improved survival and end-of-life outcomes for patients with pancreatic cancer: a cohort study

Table 4 Sensitivity analysis of multivariable regression analyses excluding patients without stage data ^a

	Mortality	Healthcare encounters per 30 days	Palliative care visits per 30 days	Institutional death^b	Hospitalization within last 30 days of life	Chemotherapy within last 30 days of life
	HR (95% CI)	AMD (95% CI)	AMD (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
No cancer-directed therapy (N = 1479)	Referent	Referent	Referent	Referent	Referent	Referent
Radiation (N = 299)	0.97 (0.85 to 1.10)	−0.09 (−0.76 to 0.57)	−0.47 (−1.11 to 0.17	0.91 (0.71 to 1.17)	0.92 (0.72 to 1.19)	2.24 (1.02 to 4.92)
Chemotherapy (N = 1701)	0.49 (0.45 to 0.53)	−2.96 (−3.35 to −2.56)	−1.50 (− 1.87 to − 1.12)	0.93 (0.80 to 1.08)	0.92 (0.79 to 1.06)	21.66 (13.94 to 33.67)
Surgery alone (N = 415)	0.42 (0.37 to 0.48)	−2.76 (−3.45 to − 2.07)	−1.23 (− 1.88 to − 0.57)	0.83 (0.64 to 1.07)	0.74 (0.57 to 0.96)	7.86 (4.30 to 14.36)
Surgery and chemotherapy (N = 406)	0.37 (0.32 to 0.42)	−2.55 (− 3.27 to − 1.84)	−1.01 (− 1.69 to − 0.33)	0.70 (0.54 to 0.92)	0.77 (0.59 to 1.01)	13.51 (7.62 to 23.97)

HR hazard ratio, AMD absolute mean difference, CI confidence interval, OR odds ratio
^aCox regression used to model survival, linear regression used to model healthcare encounters and palliative care visits per 30 days, and logistic regression to model institutional death, hospitalization within 30 days of life, and chemotherapy within 30 days of life
^bIncludes emergency department, acute care ward, intensive care unit, complex continuing care and rehabilitation facility, long-term care

ISSN: 1471-2407