Nst Suspicious Type Of High-risk Pregnancy Discrimination Based On Partial Functional Logistic Models

Non-stress test(NST) is a measurement for electronic fetal heart monitoring, which is a simple, non-invasive test performed in pregnancies over 26 or 28 weeks gestation. To understand fetal reserve capacity, NST tests fetal heart rate without posing any physical risks to fetus or gravida. This NST test can indicate whether the fetus is not receiving enough oxygen because of placental or umbilical cord problems. Situations described as bellow could be defined as atypical: 1) Presence of 1 or less discernible fetal movement with fetal heart rate accelerations within a 20-min period; 2) fetal heart rate acceleration less than 15 bpm and less 15 seconds; 3)reduced baseline variability; 4)abnormal fetal heart rate baseline(>160bpm or <120bpm); 5)spontaneous variable deceleration.In this study, we have determined some possible prognostic factors which could help to identify the potential high-risk pregnancy. Discovering the association between these factors and high-risk pregnancy does help to recognize the high-risk pregnancy before delivery, which makes it possible to be well prepared. The data was the clinical records from NST-atypical gravida. Meanwhile, we recorded the heart rate of both fetus and gravida in a long time-window(2hrs~24hrs). The main work in this study is processing, analyzing, modeling these clinical data, determining valuable factors and building predicting model.Nowadays, short-time NST test could not be good enough in identifying high-risk pregnancy. The study on fetus heart rate long-time monitoring will surely be a breaking point of researchers. This is an unsolved but urgent issue that building up a decision rule to identify high-risk pregnancy basing on clinical records. On medical economics, this model could significantly decrease the physical examination frequency of NST atypical gravida, which could help to assign medical resources better. This study has not only theoretical value on statistics, but also clinical value on medical practice.We have deeply analyzed and mined the collected clinical data, and built up a partial functional logistic model. To trace the relationship between clinical factors and high-risk pregnancy statistically, we analyzed the data using a logistic model. Moreover, we did a descriptive data analysis on the factors. Here, spearman correlation was implemented to determine the input predictors in the final model: comparing the relative weight of endpoints which measured the contribution to thehigh-risk pregnancy(response). While analyzing the functional data, we proposed data-preprocessing and functional principal component analysis. And then, a partial functional logistic model with estimated parameters revealed the connection between selected clinical factors and high-risk pregnancy. The objective of this study is to determine a rule to identify NST atypical high-risk pregnancy primarily with clinical records and maternal-fetal heart rate long-time monitoring data.