Maternal mortality rates (or ratios) are widely used as one indicator
of the quality of medical care. There has been concern that conclusion
s may be drawn from unreliable data. A survey of practices in European
countries confirms that there are differences in interpretation of de
finitions and evidence of significant under-reporting in official retu
rns of vital statistics. Actual and potential sources of error are ide
ntified and a need for greater precision and uniformity of data collec
tion is established.