Spoilt for choice: implications of using alternative methods of costing hospital episode statistics

Health Econ. 2012 Oct;21(10):1201-16. doi: 10.1002/hec.1785. Epub 2011 Sep 9.


In the absence of a 'gold standard' to estimate the economic burden of disease, a decision about the most appropriate costing method is required. Researchers have employed various methods to cost hospital stays, including per diem or diagnosis-related group (DRG)-based costs. Alternative methods differ in data collection and costing methodology. Using data from Scotland as an illustrative example, costing methods are compared, highlighting the wider implications for other countries with a publicly financed healthcare system. Five methods are compared using longitudinal data including baseline survey data (Midspan) linked to acute hospital admissions. Cost variables are derived using two forms of DRG-type costs, costs per diem, costs per episode-using a novel approach that distinguishes between variable and fixed costs and incorporates individual length of stay (LOS), and costs per episode using national average LOS. Cost estimates are generated using generalised linear model regression. Descriptive analysis shows substantial variation between costing methods. Differences found in regression analyses highlight the magnitude of variation in cost estimates for subgroups of the sample population. This paper emphasises that any inference made from econometric modelling of costs, where the marginal effect of explanatory variables is assessed, is substantially influenced by the costing method.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Age Factors
  • Aged
  • Costs and Cost Analysis
  • Data Collection / methods*
  • Diagnosis-Related Groups / economics
  • Female
  • Hospital Charges / statistics & numerical data*
  • Hospitalization / economics*
  • Humans
  • Length of Stay / economics
  • Longitudinal Studies
  • Male
  • Middle Aged
  • Models, Economic*
  • Regression Analysis
  • Scotland
  • Sex Factors
  • Socioeconomic Factors