STOCHASTIC MODELING OF FETAL HEART RATE VARIABILITY WITH APPLICATIONS TO CLINICAL RISK-ASSESSMENT AND DETECTION OF FETAL RESPIRATORY MOVEMENTS-PRELIMINARY RESULTS

Abstract

Several measures of fetal heart rate variability have been proposed (Laros et al. [27]). These simple measures were intended to provide a diagnostic tool to the clinician, when managing high risk obstetric patients. However, none of these measures, even when applied to highest quality data in research environments, could serve predictive purposes. Under clinical conditions, where a variety of recording techniques with data of rather limited quality are the rule, no value may be expected from those measures. As an alternative, we started out with a systematic stochastic modeling approach in order to extract most of the information contained in observed fetal heart rate data. For this purpose we developed a stochastic nonlinear model which permits one to give a statistically fair representation of the data. Typical clinical data from a variety of modem recording techniques are easily analyzed and yield consistent results. Examples of the power of this stochastic modeling approach are presented.