Table 3

Difference in overall Medicare spending in the last 30 days of life by primary outpatient oncologist type

Cohort	Model 1: Simple				Model 2: Multivariable
Cohort	n	Medicare spending ($)	95% CI	P value	Medicare spending ($)	95% CI	P value
Entire cohort*
Medical oncologist	7705	76 776	73 652 to 79 900	0.263	74 849	72 227 to 77 471	0.004
Gynecologic oncologist	4484	80 548	74 727 to 86 370	0.263	83 859	78 673 to 89 045	0.004
Propensity score-matched cohort, with replacement†
Medical oncologist	2768	75 769	70 650 to 80 888	0.227	73 192	68 451 to 77 933	0.013
Gynecologic oncologist	4484	80 548	74 725 to 86 371	0.227	82 139	77 246 to 87 032	0.013
Propensity score-matched cohort, without replacement‡
Medical oncologist	4484	75 678	71 675 to 79 681	0.177	73 629	70 184 to 77 074	0.003
Gynecologic oncologist	4484	80 548	74 726 to 86 370	0.177	82 597	77 867 to 87 327	0.003

*Covariates included in the multivariable regression models: age at death, race, ethnicity, marital status, median income of residential zip code at death, percent of people with less than a high school education in the residential zip code at death, SEER registry at death, residential urban status at death, year of diagnosis, year of death, cancer site, cause of death, stage at diagnosis, Medicare/Medicaid dual eligibility at death, and CCI at death.
†Covariates included in the multivariable regression models: SEER registry at death, cause of death, and CCI at death.
‡Covariates included in the multivariable regression models: SEER registry at death, cause of death, and year of death.
CCI, Charlson Comorbidity Index; CI, confidence interval; SEER, Surveillance, Epidemiology and End Results .