The foetal heart rate (FHR) response to uterine contractions is crucial to detect foetal distress by electronic FHR monitoring during labour. We are developing a new automated system (OxSys) for decision support in labour, using the Oxford database of intrapartum FHR records. We describe here a novel technique for automated detection of uterus contractions. In addition, we present a comparison of the new method with four other computerised approaches. During training, OxSys achieved sensitivity above 95% and positive predictive value (PPV) of up to 90% for traces of good quality. During testing, OxSys achieved sensitivity = 87% and PPV = 75%. For comparison, a second clinical expert obtained sensitivity = 93% and PPV = 80%, and all other computerised approaches achieved lower values. It was concluded that the proposed method can be employed with confidence in our study on foetal health assessment in labour and future OxSys development.

Original publication

DOI

10.1007/s11517-009-0538-9

Type

Journal article

Journal

Medical & biological engineering & computing

Publication Date

12/2009

Volume

47

Pages

1315 - 1320

Addresses

Nuffield Department of Obstetrics and Gynaecology, University of Oxford, The John Radcliffe Hospital, Oxford, UK. Antoniya.Georgieva@obs-gyn.ox.ac.uk

Keywords

Humans, Fetal Distress, Cardiotocography, Sensitivity and Specificity, Predictive Value of Tests, Pregnancy, Uterine Contraction, Algorithms, Decision Support Techniques, Signal Processing, Computer-Assisted, Female