Monitoring pregnancy

A model for consecutive multivariate observations of pregnancy associated qualtities is recursively formulated, incorporating the possibility of sudden perturbations in the parameters determining the pattern. The model constitutes the basis of a monitoring system applying the Kalman filter as suggested by Harrison and Stevens. The monitoring system is, as the model, recursively formulated and provides probabilities for a range of patterns deviating from the normal. Principles for choice of initial values and prior distributions connected to the monitoring system is discussed. Finally the paper relates the monitoring system to methods of monitoring used by the clinicians today.