Study On The Trajectory And Predictive Model Of Prenatal Psychological Disorder Of Pregnant Women

Background: During pregnancy,pregnant women are prone to produce negative emotions such as stress,anxiety and depression,which are harmful to the physical and mental health of pregnant women and fetuses.A few studies have shown that the symptoms of stress,anxiety and depression of pregnant women may change with the gestational weeks,and these trends have subgroups.At present,there are few studies on the longitudinal change trend of pregnant women’s psychology in China,and there is no report about the heterogeneity of the prenatal psychological change trajectory of women in China’s mainland.Therefore,this study uses a longitudinal evaluation of the psychological disorders of women in Chongqing to explore whether there is heterogeneity in the trends of prenatal stress,anxiety,and depression.The characteristics of the trajectory subgroup provide information for pregnant women’s prenatal psychological intervention work and the prediction model of psychological disorder trajectory.Methods: The data comes from a pregnancy longitudinal follow-up study conducted in Chongqing,China.The subjects were 916 women who completed the longitudinal psychological assessment from January 2018 to November 2019,using the Pregnancy Stress Scale(PPS)and Hamilton Anxiety Scale(HAMA)and Self-Rating Depression Scale(SDS)evaluate the psychology of pregnant women.Use Mplus7.0 to construct a growth mixed model(GMM)to explore the heterogeneity of pregnant women’s psychological change trends and potential trajectory group characteristics,to establish multiple logistic regressions with trajectory subgroups as dependent variables to explore the influencing factors of trajectory subgroups.Finally,random forest,support vector machine and logistic regression are used to establish a predictive model of pregnant women’s mental trajectory.Results: From the first trimester to the third trimester,the incidence of pregnancy stress in pregnant women was 92.25%,90.39% and 88.65%;the incidence of anxiety in pregnant women was 14.08%,11.24% and 10.04%;the incidence of depression in pregnant women was 4.80%,3.71 % And2.73%.Before childbirth,the incidence of stress,anxiety and depression all showed a downward trend.Named after the characteristics of the degree distribution and change trend of the trajectory group,we have identified 3pressure trajectory groups: low-level-stable pressure group(81.9%),highlevel-declining pressure group(2.6%)and medium-level-increasing pressure group Group(15.5%),the risk factors related to the stress trajectory group are high education level,first childbirth,school education for pregnant women,early vaginal bleeding,low family care and low level of social support;3 anxiety trajectory groups: no obvious anxiety Group(73.0%),low-level-stable anxiety group(23.6%)and medium-level-increasing anxiety group(3.4%).The risk factors associated with the anxiety trajectory group are living in a rural area,having a history of previous diseases,and taking early Abnormal drugs,keeping pets at home,low family care and low level of social support;4 depression trajectory groups: no depression group(81.8%),very low level-increasing depression group(10.8%),low leveldecreasing depression group(5.5%)and the low-level-increasing depression group(2.0%).The risk factors associated with the depression trajectory group are pet keeping at home,poor independence,low family care,and low level of social support.Low family care and low level of social support are common risk factors for prenatal stress,anxiety and depression trajectory group.The SMOTE oversampling method is used to balance the training set.In the prediction model,the random forest model has the strongest ability to identify positive samples(Se=63.89%)and comprehensive diagnostic ability(accuracy=76.98%).Logistic regression and svm Radial models are more sensitive The difference is less than 50%,and the sensitivity of the Logistic regression model that is not trained with the balanced training set is only10.71%.Conclusion: The occurrence of stress,anxiety and depression of pregnant women is dynamic,and the trend of change is heterogeneous.Our research results verify the necessity of multiple evaluations for pregnant women from the first trimester to the third trimester,and provide scientific support for medical staff to identify high-risk groups of female psychological disorders during pregnancy,so that medical staff can take timely and effective psychological intervention measures.At the same time,we found that after balancing the imbalanced data,the prediction model effect is better.