State-Level Variations in Hospital Expenditures: An Application of Spatial Regression

Abstract

Hospitalizations have reemerged as a priority for the United States healthcare system, policy makers, and research communities, due to their large share of total expenditures and morbidity and mortality burden on patient populations. Estimates of the proportion of total healthcare expenditures due to hospitalizations vary from 30-38%. While there are many reasons for hospitalization, an estimated 25% to 40% of all hospitalizations are for treating alcohol-related complications. Furthermore, it is plausible that the hospital spending in one state can affect the hospital expenditures of the neighboring states because 1) hospital expenditures of one state may be influenced by hospital prices, practices and policies of the neighboring states, and 2) presence of hospital facilities in a Hospital Referral Region (HRR) which serve patients across state lines. Therefore, using a Spatial Durbin Fixed Effect Model, this paper examines the state-level variations in hospital expenditures from 2000 through 2009, extracted from publicly available data files.

Results highlighted that rate of binge drinking, hospitals and beds, unemployment rate, African-Americans, active physicians and state GDP had positive impacts on its neighboring states. Moreover, male population, Hispanics and un-insurance rate had negative impacts on own state’s hospital costs but positive impacts on its bordering states.

Keywords: hospital expenditures, spatial regression, state level, binge drinking

JEL Classification: I100, H750, I180, C23