Date of Award
Doctor of Philosophy
Ron A. Cisler
Mary Eapen, Susan E. Cashin, Mark V. Johnston, Scott J. Adams
Acute Leukemia, All Payer Claims, Healthcare Cost, Health Economics, Insurance Cost
Understanding cost predictors of low incidence high cost cancers is increasingly important as the U.S. attempts to control health care costs. Acute myeloid and acute lymphoblastic leukemia are hematologic cancers requiring high cost care.
Using Anderson's model of health care utilization this study explores the influence of patient and community factors on health care costs. Insurance claims cost data obtained from the Wisconsin Health Information Organization provided a sample of 837 acute leukemia patients from April, 2009 and June, 2011. Total, ancillary, inpatient, outpatient, pharmacy and professional services costs were analyzed using a GLM gamma log link regression model to identify cost predictors. The added influence of patient and community enabling factors over patient characteristics and need for services was analyzed with hierarchical regression.
Study findings include: (1) Predisposing characteristics of acute leukemia patients may not follow the commonly reported direction of cost where higher cost was associated with older age and female gender. Instead their costs are expected to be higher in younger, male patients; (2) As expected, treatment with hematopoietic stem cell transplant (HCT) and increased severity of disease represent significant cost drivers and strongly influence higher costs; (3) Community enabling resources influence cost where academic medical centers are associated with higher cost and providers located in areas of higher poverty are associated with lower cost; and (4) Costs related to different service types, i.e. inpatient, outpatient, etc., may not follow the same trends and result in important differences in findings. While this creates complexity in assessing cost drivers it can provide the opportunity for cost interpretation and decision making specific to service type.
Implications of study findings support the need for healthcare service research of rare diseases; further exploration of the relationship between treatment choice and cost as well as treatment disparities between providers and their locations; and the importance of clarity in service type cost.
Future research opportunities would include linking cost data to clinical outcomes data; expanding the cost dataset longitudinally to accommodate more patient records along with a longer timeline of data for each; and sub analyses of potential interactions between variables.
Steinert, Patricia Ann, "Predictors of Healthcare Cost in a Wisconsin Acute Leukemia Population: Utilization of a State-Level All Payer Claims Database" (2012). Theses and Dissertations. 205.