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

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