Abstract
Background/Aims Estimates of the nation’s costs of cancer care are largely based on the SEER-Medicare datalink supplemented by modeling to estimate costs for non-aged patients. The SEER-Medicare data, however, are missing data on HMO (Medicare Advantage) enrollees with cancer. To address this shortfall, this study estimated the incremental medical care resources used to diagnose and treat cancer within integrated delivery systems stratified by phases of cancer care and patient age—before and after age 65.
Methods This study was conducted in four large group-model HMOs: Group Health Cooperative, Henry Ford Health System, Kaiser Permanente Colorado, and Kaiser Permanente Northwest. Eligible health plan members were enrolled in one of the participating health plans for at least 30 days between January 1, 2000, and December 31, 2008 and received a cancer diagnosis, documented in the health plan’s tumor registry, prior to December 31, 2008. Individuals selected as controls met enrollment criteria and had no evidence of cancer before January 1, 2009 in the tumor registry. Controls were matched by age, gender, and years of health plan eligibility on a 5:1 ratio to cancer patients. Pseudo-diagnosis dates were assigned to non-cancer patients. Four phases of cancer care were defined: Prodromal/Diagnosis Phase—12 months prior to cancer diagnosis; Treatment Phase—12 months after cancer diagnosis; Surveillance Phase—period between Treatment Phase and End-of-Life Phase; and End-of-Life Phase—12 months prior to date of death (regardless of cause of death). Utilization data were extracted for all patients and controls for 2000–2008. Utilization measures included hospitalizations, inpatient days, doctor office visits, laboratory tests, imaging procedures, and medication dispensings. Average incremental monthly utilization rates were computed for each care phase for all cancer patients and for four specific cancer sites—breast, prostate, lung, and colon/rectum.
Results Incremental average monthly resource-use rates were highest in the treatment and end-of-life phases, and lowest in the diagnostic and surveillance phases. The duration of the surveillance phase was shorter for cancers diagnosed at later stages.
Conclusions Age at diagnosis had a non-linear impact on incremental costs of cancer care.
