Efficiency of the German Hospital Sector - Evidence After the German Reunification

Posted: 20 Jun 2007

See all articles by Gunnar Dittrich

Gunnar Dittrich

Dresden University of Technology

Alexander Karmann

Technische Universitaet Dresden

Abstract

Rationale: This study focusses on the question of efficiency in the regulated German hospital sector under the influence of strong cost constraints and restructuring processes in the aftermath of German reunification.

Objectives: The objectives of this paper are to measure and to compare efficiency scores of hospitals and their departments in the selected rural German Bundeslaender. In this context it is an issue to compare 'old' and 'new' (former East Germany) Bundeslaender and their structure to find some insights to restructure the hospital sector in the future.

Methodology: To define economic pressure on the hospitals we start by deriving a goal function from the laws regarding the hospital sector. Hypotheses are drawn from selected efficiency studies and the descriptive statistics. The efficiency of the selected 300 hospitals (about 900 observations over three years) is measured with Data Envelopment Analysis (DEA). As inputs, different categories of personnel, material costs, beds and hospital days are chosen and as outputs, five main categories of cases plus a rest category are taken. We use a three-stage approach:

First stage: The efficiency scores are calculated for hospitals as a whole and, in a second part, for the two main departments (internal medicine, surgery). Through this homogenization, we obtain results which are more stable. Further objects of analysis are scale efficiency, development over time periods (year-by-year means, Malmquist) and efficiency distribution. Three models are chosen to find some insights about the structure of cost components and to capture given outsourcing effects. We perform outlier and sensitivity analyses (variations of input/ output, discretionary/ non-discretionary variables).

Second stage: The efficiency scores are now regressed on environmental variables (state, size, hierarchies, ownership, competition, year). Some of these variables are employed the first time for a German dataset. Different regression models are estimated to find some evidence about the 'right' way of regression (pooled/ panel). First we estimate the well known approaches OLS (with superefficiency), Tobit model and than truncated regression (in notice of the criticism to the distribution of efficiency scores) and solve the problem of serial correlation of the efficiency scores and the correlation between the residuals following a suggested algorithm by Simar/ Wilson (2007).

Third stage: The significant environmental variables found in the second stage are integrated as non-discretionary inputs/ outputs (depending on the sign of the dependency) into the DEA model for which new efficiency scores are calculated. This leads to insights about the true managerial inefficiency.

Results/ Conclusion: The analysis shows that efficiency in the ¿new¿ German Bundeslaender is higher than in the 'old' ones. This gives some evidence for the theory of bureaucracy especially in the 'old' German states. Furthermore, we have new evidence about scale effects in the hospital sector and the impact of environmental variables. This helps to give additional insight (w. r. t. scale efficiency et al.) for future hospital planning in Germany as well as for the German compulsory health insurance.

Suggested Citation

Dittrich, Gunnar and Karmann, Alexander, Efficiency of the German Hospital Sector - Evidence After the German Reunification. iHEA 2007 6th World Congress: Explorations in Health Economics Paper, Available at SSRN: https://ssrn.com/abstract=994971

Gunnar Dittrich (Contact Author)

Dresden University of Technology ( email )

Einsteinstrasse 3
Dresden, 01062
Germany

Alexander Karmann

Technische Universitaet Dresden ( email )

Muenchner Platz 1- 3
Dresden, 01069
Germany

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
849
PlumX Metrics