[Caesarean section in german hospitals: validity of hospital quality report data for monitoring C-section rates]

OBJECTIVE

It is not known if "hospital quality reports" (HQR) document Caesarean (C-) section rates at the hospital level accurately enough for use as a reliable data source when it comes to explaining regional variations of C-sections in Germany by factors at the hospital level. We aimed to answer this question using HQR from hospitals in Baden-Württemberg as data source.

METHOD

Diagnostic and procedure codes from HQR for the year 2008 (HQRdata), were used to calculate numbers of births, numbers of C-sections, and rates of births by C-section (CSR) for 94 of 97 hospitals in Baden-Württemberg. These numbers were compared to internal hospital (IH) data delivered upon request by 80 of 97 hospitals and stemming from vital statistics, birth registry forms, or external quality assurance datasets.

RESULTS

There was no difference in the number of births between HQR data and IH data, but the number of C-sections and the CSR differed significantly (p<0.05; Wilcoxon rank sum test). CSR calculated using HQR data was 4.9 ± 17.9% higher than CSR from IH data (absolute difference 1.5 ± 5.8%). The correlation between the 2 data sources was moderate (r=0.73). Only 55% of the variance in IH data-based CSR was explained by HQR data. The proportion between highest and lowest CSR in hospitals in Baden-Württemberg was 4.9 for HQR data and 3.6 for IH data.

CONCLUSION

There are significant and relevant differences between C-section rates based on ei-ther HQR or IH data. This questions routine data from HQR for 2008 as a reliable data source for research work.