OBJECTIVES: Economic models of novel cancer treatments rarely use real-world data to estimate end-of-life (EOL) costs. Obtaining improved estimates may lead to important cost offsets for novel interventions that delay progression, or are curative, and extend life. This study assessed total direct medical costs in the 6 months before death (MBD) for patients diagnosed with breast cancer (BC), prostate cancer (PC), or lung cancer (LC).

METHODS: This retrospective cohort study used Medicare 5% data between 04/01/2012 and 03/31/2016. Patients with BC, PC, or LC were placed into cohorts and were identified based on: (1) evidence of death between 04/01/2013 and 03/31/2016, (2) ≥1 encounter with a diagnosis of BC, PC, or LC in any setting in the 6 MBD, (3) age ≥65 years on the date of death, (4) continuous enrollment in Medicare with medical and hospital benefits in the 6 MBD, and (5) no days of commercial enrollment. Outcome measures included mean all-cause direct medical costs (Medicare amount paid), overall and by setting of care (inpatient hospital, skilled nursing facility [SNF], hospice, home health [HH], other settings), in the 6 MBD.

RESULTS: Among BC patients, the mean (SD) all-cause direct medical costs in the 6 MBD were $39,963 ($38,564). For LC and PC, mean (SD) costs in the 6 MBD totaled $41,618 ($38,074) and $42,603 ($42,543), respectively. Of the total direct medical costs, more than half of the costs across each cohort (58-61%) were attributable to inpatient hospitalizations. Other settings, SNF, hospice, and HH costs amounted to 13-17%, 9-13%, 8-10%, and 4-6% of total costs, respectively, across the cohorts.

CONCLUSIONS: Patients diagnosed with solid tumor cancers face significant burden with respect to EOL costs which should be incorporated in economic models to estimate the value and potential cost offsets of novel therapies.