A novel step towards machine diagnosis of fetal status in-utero: Calculation of baseline variability

Cardiotocograph is a non-invasive technique whereby fetal heart rate and uterine pressure against time is recorded. While the usefulness of this procedure is beyond question, the specificity and predictive accuracy has of late been subject to a lot of controversy. The main reason for its poor specificity has arisen due to dichotomy of opinion when examined by different clinicians. The issue has further been complicated due to different guidelines developed by various obstetrician bodies. Baseline and variability are two main components of a trace upon which a wholesome impression about the fetus is made by visual inspection. An automated interpretation system is needed to decipher the early signs of fetal hypoxia thus avoiding further compromise. Our objective is to develop a reliable algorithm to calculate variability in the light of NICHD guidelines. Variability of baseline has been calculated in 3 consecutive windows of 10 min each. For validation, results the variability of baseline have been compared with observations made by 3 clinicians. Various statistical methods such as confidence interval, paired sample t-test and Bland-Altman approach were used to measure the accuracy of the result. The results obtained were found to be comparable with clinician's observation. So, it can be incorporated into CTG machine for self interpretation if the results in a larger study are found satisfactory.