| Author/Editor | Zaletel-Kragelj, Lijana | |
| Title | Analiza učinkovitosti multiple regresije, diskriminantne analize in logistične regresije pri identifikaciji in vrednotenju napovedanih dejavnikov prezgodnjega poroda | |
| Type | monografija | |
| Place | Ljubljana | |
| Publisher | Medicinska fakulteta | |
| Publication year | 1997 | |
| Volume | str. 158 | |
| Language | slo | |
| Abstract | Objective. In perinatology one of the most important problems is a problem of pre-term labour and delivery as it is a cause of many neonatal complications with late sequellae. There is a lack of methods for recognition of pregnancies at risk for pre-term labour. Therefore several systems for prediction of pre-term labour have been developed and compared to each other. Methods. In 22902 primigravid-primiparous women, 6757 multigravid-primiparous women and 30727 multiparous women the following factors were assessed: maternal age, marital status, education, smoking habits, social status, number of previous pregnancies, number of previous labours, spontaneous and induced abortions, past pre-term labours, caesarean sections, extrauterine pregnancies, history of dead born infants, maternal history, gynecological and obstetrical history, observed pregnancy pathology and drugs used. Multiple linear regression, discriminant analysis and multiple logistic regression were used as methods for assessing the relationship between these factors and pre-term labour. Results. The percent of true positive classifications (nosological sensitivity) of models was in the group of primigravid-primiparous women 36.1% in linear regresison, 35.0% in discriminant analysis and 37.0% in logistic regression; in the group of multigravid-primiparous women these percentages were as follows 37.8% in linear regression, 37.5% in discriminant analysis and 40.4% in logistic regression, and in the group of multiparous women they were 45.6% in linear regression, 45.5% in discriminant analysis and 47.7% in logistic regression. The best performance was shown by logistic regression. (Abstract truncated at 2000 characters.) | |
| Descriptors | LABOR, PREMATURE DISCRIMINANT ANALYSIS REGRESSION ANALYSIS LOGISTIC MODELS PREGNANCY RISK FACTORS |