The effects of DRG-based financing on hospital performance: productivity, quality and patient selection

An analysis of productivity growth covering the period 1999 to 2014 indicates an increase in productivity of 25 % over this 15 year period. Productivity growth was particularly strong in the period following the hospital reform in 2002. This period has also been charceterized by a reduction in the number of hospitals, both due to closures and organizational mergers. Despite this reduction measures of acsle efficiency remain stable. Estimates of optimal scale, however, shows this to increase, while the majority of hospitals are found to have an actual size that is larger that estimated optimal size. Changes in coding behaviour, so called DRG-creep, may influence estimated productivity. Analysis, however, shows that although there are indications of DRG-crecreep in the period 1998-2008, the overall respons to changes in "profitability" of upcoding has been quite small.

Despite widespread use little is known about the relationship between activity based financing and hospital performance. Studies of the Norwegian reform in 1997 indicate an initial boost in productivity as do similar studies of the English pay for perform ance system. In a Norwegian (and Nordic) context analysis of long term effects are however scarce as are analysis of the relationship between payment systems, quality and patient selection. We will expand on existing data from the Nordic countries to esta blish a data set describing activity and costs for the period 2000 to 2011/12. With data covering a period of 12 years we will be able to provide measures of long term productivity growth in the Norwegian hospital sector, and relate this to the extent of and fluctuations in the degree of activity based financing. Specifically we will also look for associations between activity based financing and quality as well as between activity based financing and patient selection and DRG-creep. The project is linke d into an existing EU FP7 project in which the project group has the responsibility for one WP. It will also draw substantially on an international advisory group consisting of researchers from countries with similar systems - i.e. UK, Spain, Portugal as well as the other Nordic countries. In the analysis Finland, Sweden and Denmark will act as "control groups" to the Norwegian hospitals. The project is ambitions in the use of methodology - productivity measures will be obtained both by using non-parametr ic (DEA) and parametric methods (SFA). Analysis of associations between hospital payment and productivity/quality/selection will utilize multilevel regression techniques. The project will employ two PhD candidates as well as senior personnel.