OBJECTIVES: In-patient treatment of community-acquired pneumonia (CAP) is a generally non-invasive therapy with financial burden depending on length of hospitalization. In 2004, the concept of diagnosis-related groups (DRG) was implemented in Germany. Objective of this study was to analyse the effect of DRG implementation on length of hospital stay. METHODS: Open, prospective observational study conducted in German hospitals before and during DRG-implementation. Between October 2002 and July 2003 patients with CAP were enrolled (cohort A). Between January and October 2005 a second investigation was conducted (cohort B). No intervention in physicians' treatment decisions was taken. Diagnostic measures, drug application, non-medical therapy, nursing time and length of hospital stay were documented by the treating physicians. RESULTS: In cohort A 319 patients were documented in 9 hospitals, in cohort B 322 patients in 14 hospitals. Cohort A and B showed no significant differences in demographic data, clinical condition and severity of CAP classified according to Fine et al. 88.7 % (A) and 92.9 % (B) of patients were discharged from hospital due to successful treatment. 11.3 % (A) and 7.1 % (B) of patients died. Average length of stay was 10.67 (A) and 10.05 days (B) in peripheral ward (p=0.229), and 0.79 (A) and 0.75 days (B) in ICU (p=0.847). Overall length of stay was 11.46 (A) and 10.80 days (B) (p=0.162). CONCLUSIONS: In this study the effect of DRG-implementation, measured by length of hospital stay, was marginal. Between 1994 and 2000 a dramatic reduction of both length of stay and number of hospital beds were observed because of a number of new laws aiming at cost reduction. Therefore, the DRG concept for CAP patients has not yet led to a significantly reduction of length of stay in the hospital sector.