[Assessing the impact of delivery unit size on neonatal survival: estimation of potentially avoidable deaths in Hessen, Germany, 1990-2000].

BACKGROUND AND OBJECTIVE

There are only few analyses from Germany on the impact of delivery unit size on neonatal outcome. The objective of this study was to evaluate the influence of delivery unit size on neonatal survival in Germany.

PATIENTS AND METHODS

Data from the perinatal birth register for Hessen for 1990-2000 comprising 640554 births, and the Neonatal Survey for 1989-1997 in Hessen were used. Potentially avoidable deaths were assessed according to delivery unit size and birth weight category. Additionally trend analyses and an extrapolation to potentially avoidable deaths in all of Germany were performed.

RESULTS

Compared to large delivery units, smaller ones showed higher risk adjusted mortality rates. Calculation of potentially avoidable deaths gave an estimate of 257 early neonatal deaths in 11 years. Although trend analyses revealed a decline of potentially avoidable deaths, an extrapolation of such deaths for all of Germany still yielded an estimate of more than 300 potentially avoidable deaths per year using data from 1997-2000 only.

CONCLUSION

A valid inference as to the magnitude of the observed effect remains even in the face of a very cautious interpretation of our results. During the study period of 11 years, more than 200 neonatal deaths could be attributed to the fact that births in Hessen are dispersed among many small hospitals. If this pattern of births in small units is common throughout Germany, it suggests that several hundred neonatal deaths per year may be attributed to this risk factor when extrapolating these results nationally. Further research is necessary to describe the nation-wide magnitude of this problem and to identify the role of underlying causal risk factors more precisely. Additionally policy discussions regarding structural changes in obstetrical care should be undertaken in the meantime, aimed at reducing the observed mortality rates.