Approximately 8-10% of all babies born in the USA each year deliver before 37 weeks gestation. Current high risk scoring tools which attempt to identify pregnant women at risk for preterm labor have problems with reliability and validity, and reported studies using these tools achieve only 18-44% accuracy in predicting preterm labor.
This research aims to utilize knowledge acquisition and expert system technology to develop a computerized tool which will assist nurses with preterm labor risk assessment. Machine learning algorithms will be used to extract production rules from perinatal databases. Expert consultants will evaluate and enhance the rules in the prototype expert system, which must achieve a 50% accuracy rate to be an improvement over existing tools. Enhancements should continue to improve the expert system's accuracy. The long-term goal of the proposed expert system is to decrease the incidence of preterm labor by providing health care providers with tools that improve detection, and therefore appropriate intervention, for women at risk for preterm labor.
| Woolery, L K; Grzymala-Busse, J (1995) Machine learning for development of an expert system to predict premature birth. Biomed Sci Instrum 31:29-34 |
| Woolery, L K; Grzymala-Busse, J (1994) Machine learning for an expert system to predict preterm birth risk. J Am Med Inform Assoc 1:439-46 |
