Forecast Model for the Outcome of a Pregnancy

Summary

A hospital material of 411 records of fetal deaths and 424 records of matched control births are studied. The data consist of findings observable before the delivery in order to make it possible to use analogous data in a forecast model.

The forecast model is achieved applying the Bayesian method to the data and assuming that the variables are independent of each other. The model is written in a simple form which yields a score ranging from 0 to 100. Zero implies a very bad prognosis and 100 meaning no risk factors detected. The model has the form: ${\text{W}} = \frac{{121,9}}{{1 + 0.219 * {\text{Q}}}}$, where Q = the product of the coefficients of the found risk factors.