Abstracts

Rationale: The purpose of this study was to identify hospital characteristic factors affecting length of stay by patients with epilepsy. We used the revised theoretical model of healthcare access and utilization, the Andersen Behavioral Model. Methods: This study used information on hospital discharges of patients with epilepsy (PWE) extracted from the 2004 Healthcare Utilization Project's (HCUP) Nationwide Inpatient Sample (NIS). To extract the epilepsy population we used the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) diagnosis codes. The principal objective was to analyze the effect of patient, hospital and geographic characteristics on the Length of hospital stay (LOS) by PWE. To explain the variability of LOS, we have used Inverse Gaussian Regression. Regression models incorporated the predisposing factors (age, gender), enabling factors (patient characteristics: admission source, payer information, median household income; hospital characteristics: control/ownership, geographic region, teaching status, location, bed size), and illness/need factors (admission type, number of medical diagnoses, number of medical procedures and severity measures) to identify how these variables influence the LOS for PWE admitted to hospitals within the sampling frame. Results: The data set for the analysis totaled 261, 024 weighted hospital discharges of patients with a primary diagnosis of epilepsy. Hospital characteristic factors that significantly affected LOS were: private-non-profit (voluntary) hospitals shorter LOS than public hospitals (OR: .924, [95% CI .854 - 1.000], p < 0.05); non-teaching hospitals shorter LOS than Teaching (OR: .943, [95% CI 0.896 - 0.993], p < 0.05); Rural hospitals shorter LOS than Urban (OR: .856, [95% CI .807 - .908, p < 0.001]; Region - Midwest, South and West shorter LOS than Northeast (OR: .836, .879, and .885) [95% CI .796 - .879], [95% CI .836 - .925], and [95% CI .822 - .953], all values respectively and p < 0.001; small and medium bed sized hospitals shorter LOS than large (OR: .917 and .944, [95% CI 0.872 - 0.963 and 0.909 - 0.981], respectively, and p < 0.05. Conclusions: Controlling for several indicators, non-government hospitals report shorter length of stay. Patients with epilepsy are primarily utilizing large bed sized, teaching hospitals, which tend to be located in urban areas and report longer length of stay. In order to receive treatment for complex and sometimes severe medical condition, PWE may be seeking hospitals with greater scale and service capacity, i.e., video monitoring or surgical evaluation, as examples. Policy implications may effect a reduction in disease burden for large bed sized, teaching hospitals. Possible clinical collaborations would be warranted with neurologists affiliated with non-teaching, small or medium bed sized hospitals to implement more preventative practices, i.e., routine monitoring of therapeutic antiepileptic drug (AED) levels. Collaborations would be necessary in an effort to prevent possible lengthy hospital stay, i.e., AED dose adjustment/toxicity.