Understanding Health Care Costs in a Wisconsin Acute Leukemia Population

Purpose: We investigated factors driving health care costs of patients with a diagnosis of acute myeloid and acute lymphoblastic leukemia. Methods: Standard costs identified in insurance claims data obtained from the Wisconsin Health Information Organization were used in a sample of 837 acute leukemia patients from April 2009 to June 2011. The Andersen behavioral model of health care utilization guided selection of patient and community factors expected to influence health care costs. A generalized linear model fitting gamma-distributed data with log-link technique was used to analyze cost. Results: Type of treatment received and disease severity represented significant cost drivers, and patients receiving at least some of their treatment from academic medical centers experienced higher costs. Inpatient care and pharmacy costs of patients who received treatment from providers located in areas of higher poverty experienced lower costs, raising questions of potential treatment and medical practice disparities between provider locations. Directions of study findings were not consistent between different types of services received and underscore the complexity of investigating health care cost. Conclusions: While prevalence of acute leukemia in the United States is low compared to other diseases, its extreme high cost of treatment is not well understood and potentially influences treatment decisions. Acute leukemia health care costs may not follow expected patterns; further exploration of the relationship between cost and the treatment decision, and potential treatment disparities between providers in different socioeconomic locations, is needed