AbstractsMedical & Health Science

Methodology development algorithms for processing and analysis of cardiotocograms in labor

by Γρηγόρης Συκάς




Institution: University of Patras
Department:
Year: 2014
Keywords: Cardiotocography; Foetus; 618.320 75; Καρδιοτοκογραφία; Έμβρυο
Record ID: 1153222
Full text PDF: http://hdl.handle.net/10889/7638


Abstract

To develop and validate a computerised algorithm for the interpretation of the characteristics of the fetal heart rate (FHR) in labour. Materials and methods: A computerised algorithm based on wavelet transform and Fuzzy C-means algorithm was developed to assess baseline, variability, the presence of accelerations and types of decelerations. Twenty four segments of intrapartum cardiotocographs (CTG) were interpreted using the algorithm and evaluated by seven expert observers. The results compared to assess inter observer variation and agreement between the computer and experts. Results: Inter observer agreement for FHR baseline and the presence and type of decelerations was good (Interclass correlation coefficient (ICC), 0.93, 0.93 and 0.79 respectively). Conclusions: The validation of a computerised algorithm is limited by inter observer variation. The prediction of baseline and decelerations is as good as clinical observers but is more reproducible.