OBJECTIVES: To compare survival and healthcare utilization and costs among Medicare patients with acute myeloid leukemia (AML) versus a matched cohort of Medicare patients without cancer.
METHODS: Patients aged 65 years in the Surveillance, Epidemiology, and End Results (SEER) cancer registry with a new AML diagnosis from January 1, 1997 to December 31, 2007 were identified (first diagnosis termed “index”). Patients were required to have 6 months Medicare Part A and B benefits pre-index and no managed care enrollment post-index. Patients were excluded if they had another tumor in SEER pre-index. Medicare patients without cancer were identified and matched up to 5 to 1 based on age (5 years), gender, race, geographic location, and common comorbidities. Patients were followed from index (or index of the corresponding AML patient among controls) to death or database end (i.e., December 31, 2007). Study measures included median survival and health care utilization and costs. Generalized linear models were undertaken to estimate adjusted costs.
RESULTS: A total of 6,888 selected AML patients were matched to 22,346 controls. Among AML patients and controls respectively, mean (SD) age was 78.3(7.2) and 72.7(6.7) years, median survival was 2.6 and 131.7 months, mean (SD) total follow-up costs were $90,395($104,228) and $26,900 ($41,840), and mean (SD) average monthly follow-up costs were $26,990 ($30,719) and $269 ($468). The largest proportion of costs was hospitalization-related in both cohorts (74% and 42% of total, respectively). The cost difference between cohorts was mainly attributable to hospitalizations ($56,314 difference), followed by outpatient visits ($3,382 difference) (both p0.001). AML patients and controls had approximately the same number of emergency department, outpatient hospital, and home health visits. Regression analyses found AML patients accrued $74,177 more in costs than controls (p0.001).
CONCLUSIONS: While AML patients had shorter median survival, they accrued 3 times more costs, mainly driven by hospitalizations. This indicates a substantial economic burden incurred by AML patients to Medicare.
