Development and validation of a machine learning model for prediction of cephalic dystocia

Development and validation of a machine learning model for prediction of cephalic dystocia | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Development and validation of a machine learning model for prediction of cephalic dystocia Yumei Huang, Xuerong Ran, Jinguo Zhai, Xueyan Wang, Defang Wu, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4776419/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 18 Aug, 2025 Read the published version in BMC Pregnancy and Childbirth → Version 1 posted 11 You are reading this latest preprint version Abstract Background Early detection of cephalic dystocia is challenging, and current clinical assessment tools are limited. Machine learning offers unique advantages, enabling the generation of predictive models using various types of clinical data. Our model aims to integrate objective ultrasound data with psychological and sociological characteristics and obstetric treatment data to predict the individual probability of cephalic dystocia in pregnant women. Methods We collected data from 302 pregnant women who underwent examinations and deliveries at Southern Medical University's Nanfang Hospital from January 2022 to December 2023. We utilized basic patient characteristics, foetal ultrasound parameters, maternal anthropometric data, maternal psychological measurements, and obstetric medical records to train and test the machine learning models. Our study analysed the effectiveness of three machine learning models: least absolute shrinkage and selection operator (LASSO) regression, decision tree, and random forest. The precision, accuracy, recall, and area under the receiver operating characteristic (ROC) Curve (AUC) were used to evaluate the performance of the models. Results Among the three machine learning models, the LASSO-based logistic regression model demonstrated the best predictive performance, with an AUC value of 0.833. We found that maternal ischial spine diameter, fetal biparietal diameter, fetal biophysical profile score, artificial rupture of membranes, labor analgesia, childbirth self-efficacy, and other variables were predictive factors for cephalic dystocia. Conclusions This study constructed and validated a prediction model for cephalic dystocia via three machine learning methods, which can help clinicians improve the probability of identifying pregnant women at risk for cephalic dystocia. machine learning cephalic dystocia risk assessment prediction model Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Background As of 2021, the World Health Organization (WHO) reported a global cesarean section rate of approximately 21%, with projections suggesting that it will approach 30% by 2030 [ 1 ]. This rate has been increasing annually. Dystocia is defined as a slow progression of labor, insufficient cervical dilation, and/or lack of fetal head descent [ 2 ]. It is a common indication for cesarean section. Cephalic dystocia, the most prevalent type of dystocia during childbirth, refers to difficulties in giving birth to the foetus when the head is the first presenting part [ 3 ]. In some countries, approximately 10% of full-term cephalic pregnancies are complicated during childbirth. Among these cases, 75 ~ 80% result in natural childbirth, whereas 20 ~ 25% require a cesarean section [ 4 ]. Cephalic dystocia, a multifactorial phenomenon, arises from complex interactions involving fetal head flexion, rotation, and descent [ 5 ]. It is primarily characterized by slow or prolonged labor progression, posing risks to both mothers and babies while also negatively impacting the birthing experience [ 6 ]. This situation typically arises after a woman has been in labor for some time, making it difficult to distinguish it from normal cephalic deliveries before birth [ 7 ]. The timely diagnosis and prediction of cephalic dystocia remain key challenges in obstetrics. However, there are currently a limited number of predictive methods available for cephalic dystocia. Friedman defined the normal labor process in 1954 and introduced the classic "S" curve labor graph [ 8 ]. This graph allows for the identification of prolonged or stalled labor by surpassing specified time limits, aiding clinical professionals in timely intervention. Nevertheless, Friedman’s labor graph has faced scrutiny in recent years because it does not fully align with real-world clinical scenarios [ 9 ]. At present, digital vaginal examination (DVE) is the standard method for assessing labour progress. However, the results depend on the clinician's experience and subjective judgement, with the possibility of interexaminer error[ 10 ]. Additionally, it is an invasive procedure that carries a risk of infection [ 11 ]. The International Society of Ultrasound in Obstetrics and Gynecology has issued guidelines for the use of ultrasound during childbirth [ 12 ]. Ultrasound imaging has been proven to be more accurate and reproducible than DVE in determining foetal head position and descent into the pelvic region [ 13 ]. Nevertheless, relying solely on a single clinical examination cannot successfully predict the occurrence of head position dystocia. Maternal psychological factors and social factors may also influence labour outcomes. Integrating these factors could contribute to a more comprehensive approach to managing and predicting cephalic dystocia. Machine learning (ML), which is superior to traditional statistics, can harness extensive data to construct robust predictive models. ML has been widely applied in the fields of obstetrics and gynecology [ 14 ]. It can assess the risk of disease during pregnancy and predict the mode of childbirth. For example, it can predict the risk of preeclampsia [ 15 ], shoulder dystocia [ 16 ], and preterm birth [ 17 ]. In this study, we developed a model using machine learning methods to predict cephalic dystocia. This model is intended to supplement existing approaches and, when integrated into clinical practice, may enhance labor monitoring, assist clinicians in deciding on surgical childbirth, reduce the risk of emergency cesarean sections, and potentially improve maternal and neonatal outcomes. Methods This cross-sectional study was conducted in the Department of Obstetrics and Gynecology at Nanfang Hospital, Southern Medical University, from January 2022 to December 2023. This hospital is a university-based tertiary medical center. Informed consent was obtained from all participants. The study was approved by the Ethics Committee of the Nanfang Hospital of Southern Medical University (No: NFEC-2021-370). Patient selection The women who met the inclusion criteria were aged between 18 and 45 years, had a full-term pregnancy, were first-time mothers, carried a single live foetus in a cephalic presentation confirmed by ultrasound, and expressed a desire for a trial of vaginal childbirth. We excluded pregnant women with mental or cognitive disorders, severe pelvic deformities that preclude vaginal childbirth, abnormalities in the birth canal, severe pregnancy complications, or severe internal medical conditions. The data of a total of 320 eligible individuals were collected, of whom 18 had incomplete data. Ultimately, this study included 302 women to establish the predictive model. Data understanding and collection Cephalic dystocia refers to obstructed labour occurring in the cephalic position and requiring surgery (cesarean section or vaginal assistance) [ 18 , 19 ]. Vaginal-assisted childbirth includes the use of forceps, fetal head aspiration, and manual rotation of the fetal head. Abnormal foetal head position, such as a persistent posterior occipital position, persistent lateral occipital position, and natural childbirth after free-hand rotation of the foetal head, is also a type of cephalic dystocia [ 18 ]. In certain cases of cephalic childbirth, erroneous judgment can lead to vaginal birth resulting in stillbirth, neonatal death, or intracranial hemorrhage, subsequently causing cerebral palsy or severe intellectual disability. These cases are also classified as cephalic dystocia [ 18 ]. The information of the participating women was stored anonymously. The researchers collected a total of 53 data items, including demographic information (age, education background, occupation, monthly income, etc.) and psychological data. (1) Childbirth Self-Efficacy Inventory (CBSEI-32): This questionnaire was developed by Lowe NK [ 20 ] in 1993 and adapted into Chinese by Gao LL et al. [ 21 ]. It consists of 32 items with a total score of 320. Higher scores indicate stronger self-efficacy in late pregnancy. The Cronbach's α coefficient is 0.96. (2) Pregnancy-Related Anxiety Scale (PrAS): This scale was developed by Brunton RJ et al. [ 22 ] in 2019 and adapted into Chinese by Wu Y [ 23 ]. It consists of 32 items with a total score of 132, where a score of 75.5 or higher indicates higher anxiety levels. The Cronbach's α coefficient is 0.861. (3) Childbirth Attitude Questionnaire (CAQ): This questionnaire was developed by Lowe NK [ 24 ] in 2000 and adapted into Chinese by Zhang Ming [ 25 ]. It consists of 16 items with a total score of 64. The Cronbach's α coefficient is 0.882. In addition, the researchers extracted maternal obstetric treatment data from the obstetric electronic medical records system, including the number of pregnancies, fundal height, maternal abdominal circumference (AC), fetal ultrasound parameters, whether oxytocin was administered, whether artificial rupture of membranes (ARM) was performed, whether labor analgesia was used, and other records. Two researchers regularly reconciled the data. Data processing and statistical analysis We processed the extracted data to generate a database that facilitates algorithm selection. Missing values were imputed using predictive mean matching with the “mice” package in R version 4.3.2 ( https://www.r-project.org/ ), and log 2 transformation was applied to continuous variables with nonnormal distributions. Descriptive statistics were calculated for the baseline characteristics. Continuous variables are expressed as the mean (SD) or median (P25, P75) and were analysed using two-sample t tests or Kruskal‒Wallis tests, depending on normality. Categorical variables are presented as counts or percentages and were analysed using chi-square tests or Fisher's exact tests. Model development Using R version 4.3.2, we randomly divided 302 participants into a training set and a validation set at a 75%/25% ratio. The model was constructed based on the training set data. Given the small sample size, we employed 10-fold cross-validation with 5 repeats in each model to enhance the model's credibility. In each fold, 75% of the data were used as the training set, and 25% were used as the test set. The model's performance was evaluated on an independent validation set. The 75%/25% split ratio is commonly used in machine learning algorithms for medium or small sample sizes because it helps build robust models and allows for effective evaluation within a limited sample size [ 26 ]. We employed three machine learning approaches: LASSO-based logistic regression, decision tree, and random forest methods. In ML algorithms, LASSO regression can help select important features and reduce model complexity. Decision tree methods are easy to understand and interpret, with advantages in handling categorical variables. The random forest algorithm can provide a comprehensive measure of feature importance, offering higher prediction accuracy and helping to reduce the risk of overfitting [ 27 ]. Two iterations were conducted for the LASSO-based logistic regression model. The first model, logistic_53, included all 53 variables. The second model, logistic_20, included 20 variables that were identified based on their differences between dystocia and nondystocia women and were ranked among the top 20 in terms of P values in logistic regression analysis. The parameter α was set to 1, and λ was selected from the range 10^seq (2, -2, by -0.1). The optimal λ was determined by maximizing the AUC over 10 repeated cross-validation iterations The decision tree generated two models using 22 variables selected based on logistic regression. The first model was the CART classification tree, with complexity parameters ranging from 0.01 to 0.5 in increments of 0.01. The second model was the C4.5 tree, with parameters including boosting iterations set to 10, 20, or 30 and winnowing set to either TRUE or FALSE. The optimal parameter or parameter combination was selected based on the highest AUC after 10 repetitions of cross-validation. For the random forest model, 53 variables in the training set were used. The selection of predictor variables was based on minimizing cross-validation error, with the parameter being the number of randomly selected predictors, ranging from 4 to 10. The optimal parameters or parameter combinations were determined by selecting those with the highest AUC after 10 repeated cross-validation iterations. Ultimately, the top 20 variables were selected based on the mean decrease Gini criterion, following the principle of minimizing cross-validation error. The model's performance was evaluated using precision, recall, accuracy, p value, and the AUC with a 95% confidence interval (CI). An AUC of 0.5 to 0.9 indicates high accuracy [ 28 ]. Results Patient characteristics Between January 2022 and December 2023, a total of 302 pregnant women who met the inclusion criteria were selected for the study. Among the study sample, 62 participants (20.5%) experienced cephalic dystocia, while 240 participants (79.5%) had successful natural births. Age, height, body mass index (BMI), educational background, average monthly household income, etc., and demographic characteristicswere not significantly different between the training set and validation set (Table 1). Table 1 . Demographic characteristics of the training set ( n =227) and testing set ( n = 75) Characteristic Training set ( n =227) Testing set ( n =75) P value Age (year) 28.00 (3.68) 28.09 (3.48) 0.84 Pregnancy days at admission (day) 275.65 (6.37) 275.03 (7.08) 0.47 Maternal height (cm) 159.58 (4.82) 158.89 (5.19) 0.28 Maternal weight before pregnancy (kg) 50.37 (7.25) 51.02 (6.55) 0.48 Table 1. (continued) Characteristic Training set ( n =227) Testing set ( n =75) P value Maternal BMI before pregnancy (kg/m 2 ) 19.76 (2.59) 20.19 (2.29) 0.20 Current weight (kg) 64.26 (7.78) 65.65 (8.85) 0.19 Weight gain during pregnancy (kg) 13.88 (3.87) 14.62 (4.90) 0.38 Current BMI (kg/m 2 ) 25.21 (2.66) 25.98 (3.02) 0.05 Medical payment method (%) 0.15 Urban employee basic medical insurance 180 (79.30) 57 (76.00) Urban resident basic medical insurance 3 (1.30) 1 (1.30) completely publicly funded 1 (0.40) 3 (4.00) fully self-funded 43 (18.90) 14 (18.70) Family per capita monthly income (CNY) 0.72 ＜5000 (%) 16 (7.00) 4 (5.30) 5000～9999 (%) 91 (40.10) 26 (34.70) 10000～14999 (%) 63 (27.80) 22 (29.30) ≥15000 (%) 57 (25.10) 23 (30.70) Maternal educational background 0.47 Junior high school and below (%) 42 (18.50) 10 (13.30) High school and technical secondary school (%) 26 (11.50) 14 (18.70) College and undergraduate (%) 150 (66.10) 43 (57.30) Bachelor degree or above (%) 9 (4.00) 8 (10.70) Whether this pregnancy is planned 0.97 No (%) 67 (29.50) 22 (29.30) Yes (%) 160 (70.50) 53 (70.70) Pregnancy mode 0.08 Nature conceived (%) 211(93.00) 74 (98.70) ART (%) 16(7.00) 1(1.30) Whether continue to work in the third trimester of pregnancy 0.77 No (%) 86(37.90) 27 (36.00) Yes (%) 141(62.10) 48 (64.00) Pregnancy times, n (%) 0.51 1 (%) 164 (72.20) 59 (78.70) 2 (%) 50 (22.0) 15 (20.00) 3 (%) 11 (4.80) 1 (1.30) 4 (%) 2 (0.90) 0 (0.00) Continuous variables are expressed in mean±standard deviation (SD) or median (25th–75th percentiles). Categorical variables were expressed as frequencies (percentages). Abbreviations: BMI: body mass index, CNY: China Yuan, ART: assisted reproductive technology. LASSO-based logistic regression model d escription For the training dataset, Model Logistic_53 exhibited comparatively lower performance across various performance metrics compared to Model Logistic_20. In the testing dataset, Logistic_53 exhibited higher recall (0.967 vs 0.933) than logistic_20. However, Logistic_20 outperformed Logistic_53 in terms of precision, accuracy and AUC, as shown in Table 2. The validation performance, shown in Fig. 1, indicates a significant enhancement in prediction performance for Logistic_20. Table 2. Performance parameters of the five machine learning prediction models in the training and testing sets Data set Predictive models Precision Recall Accuracy P value AUC (95% CI ) Training set Logistic_20 0.890 0.989 0.894 1.05e-04 0.760 (0.688,0.833) Logistic_53 0.864 0.989 0.868 5.01e-06 0.697 (0.625,0.768) Random forest 1.000 1.000 1.000 NaN 1.000 (1.000,1.000) Decision tree CART 0.855 0.983 0.855 6.01e-06 0.673 (0.602,0.743) Decision tree C4.5 0.928 1.000 0.938 5.12e-04 0.851 (0.785,0.917) Testing set Logistic_20 0.933 0.933 0.893 1.00e+00 0.833 (0.713,0.953) Logistic_53 0.892 0.967 0.880 1.82e-01 0.750 (0.617,0.883) Random forest 0.892 0.967 0.880 1.82e-01 0.750 (0.617,0.883) Decision tree CART 0.919 0.950 0.893 7.24e-01 0.808 (0.682,0.935) Decision tree C4.5 0.919 0.950 0.893 7.24e-01 0.767 (0.634,0.899) Consequently, we selected Model Logistic_20 as our preferred predictive model because of its superior overall performance. The final model encompasses 18 variables, including the internodal diameter of maternal ischial bone, fetal biparietal diameter (BPD), fetal humeral diameter (HL), fetal biophysical profile score (BPP), artificial rupture of membranes, doula, and other variables (Fig. 2). The final regression equation was as follows: logit P= －0.18－0.038 × CBSEI-32 score－0.36 × Fetal biophysical profile score＋0.867 × Fetal biparietal diameter－1.38 × Fetal humeral diameter＋0.014 × Husband’s height－0.523 × Internodal diameter of maternal ischial bone－0.141 × Maternal interiliac spine diameter＋0.046 × Medical payment method＋0.058 × Pregnancy days at admission＋0.672 × Pregnancy mode－0.153 × Pregnancy times＋1.034 × Whether artificial rupture of membranes－0.122 × Whether continue to work in the third trimester of pregnancy－0.473× Whether there is accompanied labor＋0.143 × Whether induced labor with water sac＋ 0.905 × Whether there is Doula－0.224 × Whether there is labor analgesia－0.137× Whether this pregnancy is planned. Decision tree model description According to logistic regression, 22 variables were selected to construct the decision tree model. In the training dataset, Model_C4.5 outperformed Model_CART across various indicators, as described in detail in Table 2. After cross-validation, in the testing dataset, Model_CART and Model C4.5 exhibited identical precision, recall, and accuracy, with a greater AUC for Model_C4.5 than for Model CART (0.808 vs 0.767), as shown in Fig. 3. Five variables played a crucial role in the construction of the decision tree: CBSEI_32 score, BPP, ARM, doula, and HL. The resulting decision tree is depicted in Fig. 4. Random forest model description Following the principle of minimizing cross-validation error, we selected the top 20 variables based on the mean decrease Gini coefficient , where a larger "mean decrease Gini" value indicates a greater contribution of the feature to improving classification performance. Among the crucial features were the CBSEI_32 score, fetal head circumference (HC), maternal height, AC, the S/D value of umbilical blood flow, HC, and other variables, as illustrated in Fig. 5. A classification model was constructed on the basis of these selected top 20 variables, with the optimal parameter set to 8. The model achieved an AUC of 1.000 (95% CI: 1.000 to 1.000) on the training set and 0.750 (95% CI: 0.617 to 0.883) on the testing set, as shown in Supplementary Fig. 1. Overall performance comparison Model logistic_20 appears to be a promising model, as it performs well across multiple performance metrics, including precision, recall, accuracy, and AUC values, as well as small P values and relatively narrow confidence intervals. We generated AUC curves for all the models to visualize these results, as shown in Supplementary Fig. 2. Discussion Identifying cephalic dystocia prior to labor is highly challenging because of multiple influencing factors. Previous studies have reported relevant risk factors but have seldom developed predictive models [29-32]. This study utilized three machine learning methods to develop and validate a predictive model for cephalic dystocia. Fetal ultrasound parameters, maternal anthropometric measurements, psychological factors, and obstetric records were collected. The model can assist clinicians in evaluating the probability of cephalic dystocia in primiparous women at term, facilitating the early identification of high-risk women. Additionally, LASSO regression highlighted several crucial variables deserving further investigation. Ultrasound plays a crucial role in assessing fetal size and position, assisting clinicians in predicting the risk of dystocia [7]. In our study, we found that BPD, HL, and BPP were significant in the predictive model. The BPD can be used to estimate the likelihood of a foetus successfully passing through the birth canal. Shinohara S et al. [29] reported that a greater BPD might increase the risk of dystocia. In this study, a positive correlation between BPD and dystocia was also observed. A longitudinal cohort study of 2802 pregnant women revealed that fetuses of obese women had significantly longer HLs, increasing the risk of dystocia [33]. In our study, HL was negatively correlated with cephalic dystocia. A larger HL may indicate more mature fetal skeletal development, potentially aiding in withstanding pressure and facilitating passage through the birth canal during childbirth. However, excessively high HL is typically associated with macrosomia [33]. According to a multinational longitudinal study by the WHO, the HL of a foetus at 37–40 weeks is generally between 67 mm and 69 mm [34]. An HL within this normal range may be more conducive to smooth delivery. The BPP assessment helps clinicians determine whether the fetus is healthy or at risk of intrauterine hypoxia [35]. A higher BPP indicates better fetal health and is associated with a greater success rate of vaginal childbirth [36]. Similar to the negative correlation found in this study between foetal physiological score and difficult cephalic presentation, the BPP has predictive value. Furthermore, in our study, the internodal diameter of maternal ischial bone and the interiliac spine diameter were negatively correlated with cephalic dystocia. A larger internodal diameter of maternal ischial bone and interiliac spine diameter in parturients signifies a wider pelvis, which can facilitate better fetal head engagement and provide more space in the birth canal..This reduces birth canal resistance, lowering the risk of foetal head obstruction or impaction and decreasing the risk of cephalic dystocia [30]. By measuring the internodal diameter of maternal ischial bone and the interiliac spine diameter, healthcare providers can devise more personalized childbirth plans. Women with wider pelvises may be encouraged to pursue natural childbirth, while those with narrower pelvises may require early intervention, such as planned cesarean section, to ensure maternal and infant safety. ARM is typically performed after normal uterine contractions have been induced using oxytocin or mechanical methods. [37]. In our study, an ARM was identified as a risk factor for cephalic dystocia. ARM alone may impede cervical ripening, delay the labor process, and increase the risk of chorioamnionitis and umbilical cord prolapse [37]. Research has shown that using an ARM does not shorten the first or second stages of labor and tends to increase the risk of cesarean section [38]. The American College of Obstetricians and Gynecologists also advised against routine ARM solely to prevent labor prolongation [39]. Furthermore, labor analgesia was negatively correlated with cephalic dystocia in our study. Labor analgesia can reduce pain stress responses and aid uterine contractions. Moreover, it enhances the willingness of women in labor to undergo vaginal childbirth, enabling better cooperation with medical guidance and promoting a smoother labor process [40]. While labor analgesia may prolong the duration of labor, research suggests that this prolongation does not increase the incidence of postpartum hemorrhage or fetal distress [41]. Thus, healthcare providers are strongly encouraged to offer labor analgesia to eligible women, thereby decreasing the risk of cephalic dystocia. In this study, water sac induction was considered a risk factor for cephalic dystocia. Water sac induction primarily uses Foley catheters and COOK cervical ripening balloons. Obstetricians typically place balloons in the cervical canal to achieve mechanical dilation. By exerting pressure on the cervix, the balloon stimulates the release of endogenous prostaglandins, facilitating cervical ripening [42]. However, research shows that pregnant women who undergo Foley catheter balloon placement have greater infection rates (OR=1.50, 95% CI: 1.07 to 2.09) than those who undergo pharmacological induction methods, which increases the risk of dystocia [43]. Additionally, the duration and volume of catheter balloon placement may also influence the progression of labor, necessitating further exploration in future studies. This study also revealed that higher scores on the childbirth self-efficacy scale are associated with a lower likelihood of cephalic dystocia. Childbirth self-efficacy reflects women's expectations of childbirth outcomes and confidence in their abilities [20]. Research has shown a strong correlation between childbirth self-efficacy and childbirth fear [32]. Consequently, women with lower self-efficacy may be more susceptible to experiencing tension and fear, triggering catecholamine release that inhibits effective uterine contractions, thereby hindering labor progress [44]. Healthcare providers should actively engage in prenatal education and provide psychological support to pregnant women. The presence of a doula is also a significant predictor of cephalic dystocia. Experienced doulas can monitor labor progress, provide childbirth assistance, enhance maternal and infant safety, promptly identify dystocia risks, and simultaneously increase maternal childbirth self-efficacy through psychological support [45]. In this study, the presence of doula was positively correlated with dystocia. Doulas may reduce the incidence of dystocia. However, pregnant women who perceive themselves at risk for dystocia may be more likely to seek or be advised to seek doula assistance. As a result, endogeneity could affect the model, leading to estimated coefficients that may not accurately reflect the true relationship. Additionally, our sample size may be insufficient, highlighting the need for further research to explore this issue. The WHO recommends that women be supported throughout childbirth by selected peers, including midwives, doctors, doulas, spouses or other family members, and friends [46]. In this study, labor support was negatively correlated with cephalic dystocia. The maternal paternal partner is a female partner or family member. Having a companion during labor eliminates the sense of isolation, improves the childbirth experience, and contributes to a smoother birth process [47]. Research indicates that having a spouse as a companion promotes the transition of the husband's role and enhances the marital relationship [48]. However, accompanying childbirth may also have adverse effects [47]. If the companion is not well prepared, they may feel anxious when witnessing the parturient's labor. When the parturient sees negative facial expressions or body language from her companion, this may amplify her negative assessment of the situation, leading to increased anxiety and difficulty in coping with labor pain. Therefore, companions need adequate preparation and training to ensure that they can provide the best support. In this study, the method of conception was associated with cephalic dystocia. Women who conceive through assisted reproductive technology (ART) are more prone to cephalic dystocia. These women typically exhibit physiological characteristics related to infertility or comorbidities. These individuals are at increased risk of developing gestational diabetes, gestational hypertension, and abnormal amniotic fluid levels, which can affect fetal intrauterine growth and development [49]. Additionally, they may experience a heavier psychological burden during pregnancy and childbirth due to long-term ovulation induction or early pregnancy maintenance drugs [50]. As a result, these women are at greater risk of cephalic dystocia and may require emergency cesarean sections or assisted deliveries. Studies also show that the cesarean section rate is greater for women who conceive through ART than for those who conceive naturally [51]. Clinical doctors and midwives should pay special attention to the childbirth process of women who become pregnant through ART. Conversely, women who achieve pregnancy through natural conception appear to have a lower risk of cephalic dystocia in this study. This may be due to their preparation before conception, during pregnancy, and during childbirth, as well as their more favourable physiological and psychological conditions for a smooth birth. Strengths and limitations Our study has several strengths. We collected a comprehensive set of risk factor data, including foetal ultrasound parameters, maternal anthropometric measurements, mothers' psychological status, and obstetric treatment data. Our research data do not add additional costs for women. By using the LASSO regression algorithm, we removed variables that did not contribute significantly to the prediction of difficult labor, resulting in a more streamlined model with enhanced predictive performance. However, there are several limitations to our study. First, the sample size was not large enough, and external validation with a larger dataset is needed to confirm the effectiveness of the model. Additionally, potential endogeneity among independent variables may have arisen due to interactions between them, necessitating further adjustment and variable selection in a larger sample. Conclusions This study employed ML to develop a predictive model for cephalic dystocia, which demonstrated strong predictive performance. As a supplement to existing methods, the model can assist healthcare providers in the early identification of pregnant women at risk for cephalic dystocia and in making decisions about surgical delivery. Additionally, healthcare providers can focus on modifiable variables during labor, such as labor analgesia, balloon induction, labor companionship, and psychological support, to reduce the risk of dystocia. To further validate the model's results and enhance its performance, future studies should involve multicenter, large-sample prospective trials. Abbreviations LASSO: least absolute shrinkage and selection operator ROC: Receiver operating characteristic AUC: area under the ROC curve WHO: World Health Organization DVE: digital vaginal examination ML: machine learning CBSEI-32: Childbirth Self-Efficacy Inventory PrAS: Pregnancy-Related Anxiety Scale CAQ: Childbirth Attitude Questionnaire AC: abdominal circumference ARM: artificial rupture of membranes CI: confidence interval BMI: body mass index CNY: China Yuan BPD: biparietal diameter HL: humeral diameter BPP: fetal biophysical profile score HC: head circumference ART: assisted reproductive technology Declarations Ethics approval and consent to participate This study was approved by the Nanfang Hospital of Southern Medical University Ethics Committee (NFEC-2021-370). The methodology adhered to the tenets of the Declaration of Helsinki. All participants signed informed consent forms. Consent for publication Not applicable. Availability of data and materials The datasets analysed during the current study are available from the corresponding author upon reasonable request. Competing interests The authors declare no competing interests. Funding This study was supported by the Special Higher Education Program of Guangdong Provincial Education Science Planning in 2021(2021GXJK163) and the Quality Engineering Construction Project of Southern Medical University in 2021-Teaching and Research Section of Midwifery (202129). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. Authors ’ contributions YMH and XRR should be considered joint first author. YMH: Methodology, investigation, writing - original draft; XRR: Investigation, software, formal analysis, writing-original draft; JGZ: Conceptualization, project administration, supervision, writing - review & editing; XYW: Data curation, formal analysis; DFW and ZY: Investigation. Acknowledgements Not applicable. References Betran AP, Ye J, Moller AB, Souza JP, Zhang J. Trends and projections of caesarean section rates: global and regional estimates. BMJ Glob Health 2021, 6(6):e005671. Ressel GW. ACOG releases report on dystocia and augmentation of labor. Am Fam Physician 2004, 69(5):1290, 1292. Cunningham FG, Leveno KJ, Bloom SL, Dashe JS, Hoffman BL, Casey BM, Spong CY. Abnormal Labor. In . ; 2018. Caughey AB, Cahill AG, Guise JM, Rouse DJ. Safe prevention of the primary cesarean delivery. American journal of obstetrics and gynecology 2014, 210(3):179-193. Dall’Asta A, Ferretti A, Minopoli M, Ghi T. The role of the occiput-spine angle in prolonged labour and delivery outcome. Intrapartum ultrasonography for labor management: labor, delivery and puerperium 2021:319-323. Walker KF, Kibuka M, Thornton JG, Jones NW. Maternal position in the second stage of labour for women with epidural anaesthesia. The Cochrane database of systematic reviews 2018, 11(11):CD008070. Dall'Asta A, Rizzo G, Masturzo B, Di Pasquo E, Schera GBL, Morganelli G, Ramirez ZR, Maqina P, Mappa I, Parpinel G et al . Intrapartum sonographic assessment of the fetal head flexion in protracted active phase of labor and association with labor outcome: a multicenter, prospective study. American journal of obstetrics and gynecology 2021, 225(2):171.e171-171.e112. Friedman E. The graphic analysis of labor. American journal of obstetrics and gynecology 1954, 68(6):1568-1575. Zhang J, Landy HJ, Ware BD, Burkman R, Haberman S, Gregory KD, Hatjis CG, Ramirez MM, Bailit JL, Gonzalez-Quintero VH et al . Contemporary patterns of spontaneous labor with normal neonatal outcomes. Obstetrics and gynecology 2010, 116(6):1281-1287. Enomoto N, Maki S, Nii M, Yamaguchi M, Tamaishi Y, Takakura S, Magawa S, Tanaka K, Tanaka H, Kondo E et al . Accurate evaluation of the progress of delivery with transperineal ultrasound may improve vaginal delivery: a single-center retrospective study. Scientific reports 2023, 13(1):20945. Afework B, Birgoda GT, Leul H, Girma L, Bekelcho T. Frequency of digital vaginal examination and associated factors among mothers admitted to labor and delivery wards in public health facilities of Arba Minch town, southern Ethiopia. Int J Gynecol Obstet 2024, 164(3):1028-1035. Ghi T, Eggebø T, Lees C, Kalache K, Rozenberg P, Youssef A, Salomon LJ, Tutschek B. ISUOG Practice Guidelines: intrapartum ultrasound. Ultrasound in obstetrics & gynecology : the official journal of the International Society of Ultrasound in Obstetrics and Gynecology 2018, 52(1):128-139. Sherer DM, Miodovnik M, Bradley KS, Langer O. Intrapartum fetal head position I: comparison between transvaginal digital examination and transabdominal ultrasound assessment during the active stage of labor. Ultrasound in obstetrics & gynecology : the official journal of the International Society of Ultrasound in Obstetrics and Gynecology 2002, 19(3):258-263. Islam MN, Mustafina SN, Mahmud T, Khan NI. Machine learning to predict pregnancy outcomes: a systematic review, synthesizing framework and future research agenda. BMC pregnancy and childbirth 2022, 22(1):348. Marić I, Tsur A, Aghaeepour N, Montanari A, Stevenson DK, Shaw GM, Winn VD. Early prediction of preeclampsia via machine learning. American journal of obstetrics & gynecology MFM 2020, 2(2):100100. Tsur A, Batsry L, Toussia-Cohen S, Rosenstein MG, Barak O, Brezinov Y, Yoeli-Ullman R, Sivan E, Sirota M, Druzin ML et al . Development and validation of a machine-learning model for prediction of shoulder dystocia. Ultrasound in obstetrics & gynecology : the official journal of the International Society of Ultrasound in Obstetrics and Gynecology 2020, 56(4):588-596. Gao C, Osmundson S, Velez Edwards DR, Jackson GP, Malin BA, Chen Y. Deep learning predicts extreme preterm birth from electronic health records. Journal of biomedical informatics 2019, 100:103334. Ling LD. Pay attention to cephalic dystocia (article in Chinese). Chinese Journal of Practical Gynecology and Obstetrics 1994, 10(04):194-195. Qi HB. Identification and treatment of cephalic dystocia (article in Chinese). Chinese Journal of Practical Gynecology and Obstetrics 2005, 21(05):264-267. Lowe NK. Maternal confidence for labor: development of the Childbirth Self-Efficacy Inventory. Research in nursing & health 1993, 16(2):141-149. Gao LL, Ip WY, Sun K. Validation of the short form of the Chinese Childbirth Self-Efficacy Inventory in Mainland China. Research in nursing & health 2011, 34(1):49-59. Brunton RJ, Dryer R, Saliba A, Kohlhoff J. The initial development of the Pregnancy-related Anxiety Scale. Women and birth : journal of the Australian College of Midwives 2019, 32(1):e118-e130. Wu Y. Chinese version of Pregnancy-related Anxiety Scale and its reliability and validity test (article in Chinese). Master. Kunming Medical University; 2021. Lowe NK. Self-efficacy for labor and childbirth fears in nulliparous pregnant women. Journal of psychosomatic obstetrics and gynaecology 2000, 21(4):219-224. Zhang M. The research of Antenatal Fear of Childbirth and Its Associated Factors in Pregnant Women (article in Chinese). Master. Jilin University; 2016. Joseph VR. Optimal ratio for data splitting. 2022, 15(4):531-538. Ngiam KY, Khor IW. Big data and machine learning algorithms for health-care delivery. Lancet Oncol 2019, 20(5):e262-e273. Jayawant NM. Receiver Operating Characteristic Curve in Diagnostic Test Assessment. Journal of Thoracic Oncology 2010, 5(9):1315-1316. Shinohara S, Amemiya A, Takizawa M, Suzuki K. Fetal biparietal diameter as a potential risk factor for prolonged second stage of labor: A retrospective observational cohort study. PLoS One 2020, 15(10):e0239045. Jochumsen S, Hegaard HK, Rode L, Jørgensen KJ, Nathan NO. Maternal factors associated with labor dystocia in low-risk nulliparous women. A systematic review and meta-analysis. Sexual & reproductive healthcare : official journal of the Swedish Association of Midwives 2023, 36:100855. Oduro AR, Anyorikeya M, Ansah P, Oladokun S, Tei EM, Oduro-Ayeh R, Welaga P, Deh S. Birth preparedness and complications readiness among women in disadvantaged rural districts of Ghana. BMC pregnancy and childbirth 2023, 23(1):728. Dwiarini M, Chou HF, Gau ML, Liu CY. Relationship between fear of childbirth, self-efficacy, and length of labor among nulliparous women in Indonesia. Midwifery 2022, 105:103203. Zhang C, Hediger ML, Albert PS, Grewal J, Sciscione A, Grobman WA, Wing DA, Newman RB, Wapner R, D'Alton ME et al . Association of Maternal Obesity With Longitudinal Ultrasonographic Measures of Fetal Growth: Findings From the NICHD Fetal Growth Studies-Singletons. JAMA Pediatr 2018, 172(1):24-31. Kiserud T, Piaggio G, Carroli G, Widmer M, Carvalho J, Neerup Jensen L, Giordano D, Cecatti JG, Abdel Aleem H, Talegawkar SA et al . The World Health Organization Fetal Growth Charts: A Multinational Longitudinal Study of Ultrasound Biometric Measurements and Estimated Fetal Weight. PLoS medicine 2017, 14(1):e1002220. Oyelese Y, Vintzileos AM. The uses and limitations of the fetal biophysical profile. Clinics in perinatology 2011, 38(1):47-64, v-vi. Baschat AA, Galan HL, Lee W, DeVore GR, Mari G, Hobbins J, Vintzileos A, Platt LD, Manning FA. The role of the fetal biophysical profile in the management of fetal growth restriction. American journal of obstetrics and gynecology 2022, 226(4):475-486. Battarbee AN, Sandoval G, Grobman WA, Reddy UM, Tita ATN, Silver RM, El-Sayed YY, Wapner RJ, Rouse DJ, Saade GR et al . Maternal and Neonatal Outcomes Associated with Amniotomy among Nulliparous Women Undergoing Labor Induction at Term. American journal of perinatology 2021, 38(S 01):e239-e248. Smyth RM, Alldred SK, Markham C. Amniotomy for shortening spontaneous labour. The Cochrane database of systematic reviews 2013(1):CD006167. ACOG Committee Opinion No. 766. Approaches to Limit Intervention During Labor and Birth. Obstetrics and gynecology 2019, 133(2):e164-e173. Callahan EC, Lee W, Aleshi P, George RB. Modern labor epidural analgesia: implications for labor outcomes and maternal-fetal health. American journal of obstetrics and gynecology 2023, 228(5S):S1260-S1269. Anim-Somuah M, Smyth RM, Cyna AM, Cuthbert A. Epidural versus non-epidural or no analgesia for pain management in labour. The Cochrane database of systematic reviews 2018, 5(5):CD000331. Yi J, Chen L, Meng X, Chen Y. The infection, cervical and perineal lacerations in relation to postpartum hemorrhage following vaginal delivery induced by Cook balloon catheter. Archives of gynecology and obstetrics 2023. Mei-Dan E, Walfisch A, Suarez-Easton S, Hallak MJTJoM-F, Medicine N. Comparison of two mechanical devices for cervical ripening: a prospective quasi-randomized trial. 2012, 25(6):723-727. Tilden EL, Caughey AB, Lee CS, Emeis C. The Effect of Childbirth Self-Efficacy on Perinatal Outcomes. Journal of obstetric, gynecologic, and neonatal nursing : JOGNN 2016, 45(4):465-480. Fenwick J, Toohill J, Gamble J, Creedy DK, Buist A, Turkstra E, Sneddon A, Scuffham PA, Ryding EL. Effects of a midwife psycho-education intervention to reduce childbirth fear on women's birth outcomes and postpartum psychological wellbeing. BMC pregnancy and childbirth 2015, 15:284. Organization WHO. WHO recommendations on intrapartum care for a positive childbirth experience: World Health Organization; 2018. Wanyenze EW, Byamugisha JK, Tumwesigye NM, Muwanguzi PA, Nalwadda GK. A qualitative exploratory interview study on birth companion support actions for women during childbirth. BMC pregnancy and childbirth 2022, 22(1):63. Johansson M, Fenwick J, Premberg A. A meta-synthesis of fathers' experiences of their partner's labour and the birth of their baby. Midwifery 2015, 31(1):9-18. Qin J, Liu X, Sheng X, Wang H, Gao S. Assisted reproductive technology and the risk of pregnancy-related complications and adverse pregnancy outcomes in singleton pregnancies: a meta-analysis of cohort studies. Fertility and sterility 2016, 105(1):73-85.e71-76. Rozdarz KM, Flatley CJ, Kumar S. Intrapartum and neonatal outcomes in singleton pregnancies following conception by assisted reproduction techniques. Aust N Z J Obstet Gynaecol 2017, 57(6):588-592. Sha T, Yin X, Cheng W, Massey IY. Pregnancy-related complications and perinatal outcomes resulting from transfer of cryopreserved versus fresh embryos in vitro fertilization: a meta-analysis. Fertility and sterility 2018, 109(2):330-342.e339. Additional Declarations No competing interests reported. Supplementary Files SupplementaryFig1.tiff Supplementary Fig. 1 File format: tiff Title of figure: AUC for predicting the probability of cephalic dystocia via the random forest method Description of figure: The horizontal axis indicates the specificity of the risk prediction. The vertical axis indicates the sensitivity of the risk prediction. AUC: Area under the curve. SupplementaryFig2.tiff Supplementary Fig. 2 File format: tiff Title of figure: AUC for predicting the probability of cephalic dystocia via the ML models Description of figure: The horizontal axis indicates the specificity of the risk prediction. The vertical axis indicates the sensitivity of the risk prediction. AUC: Area under the curve. Cite Share Download PDF Status: Published Journal Publication published 18 Aug, 2025 Read the published version in BMC Pregnancy and Childbirth → Version 1 posted Editorial decision: Revision requested 02 Dec, 2024 Reviews received at journal 26 Nov, 2024 Reviews received at journal 11 Nov, 2024 Reviewers agreed at journal 25 Oct, 2024 Reviewers agreed at journal 25 Oct, 2024 Reviews received at journal 05 Sep, 2024 Reviewers agreed at journal 05 Sep, 2024 Reviewers invited by journal 05 Sep, 2024 Editor assigned by journal 22 Jul, 2024 Submission checks completed at journal 22 Jul, 2024 First submitted to journal 21 Jul, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4776419","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":334692720,"identity":"8bb66a75-d205-4e33-bd74-04d4e9c47639","order_by":0,"name":"Yumei Huang","email":"","orcid":"","institution":"Dongguan Maternal and Child Health Care Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yumei","middleName":"","lastName":"Huang","suffix":""},{"id":334692723,"identity":"22d36a79-dd42-4359-a763-175fe6f79ffa","order_by":1,"name":"Xuerong Ran","email":"","orcid":"","institution":"Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xuerong","middleName":"","lastName":"Ran","suffix":""},{"id":334692725,"identity":"7fe88363-3b5c-47f6-a172-6154523891e0","order_by":2,"name":"Jinguo Zhai","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7UlEQVRIie2PMQrCMBSGEwLp8tS1Ur1DoIMIYs9SBF0cBCdx0BKISw/gSTIrGVwE14KDRk8gIiguplUc246C+YbkJ7yP9wchi+UHYYjMEYI04OhoHqBarOCvwlmq0HJKFhB101CotJwoOo8aKmg5Skxuw26DIqJPSY7SjtfcX0KftONQ7JuyZ4pR3x/mFUtC4QF0KFsZpS6JUYB6ucpBL54ALrCdFuO6nJVQEiyI2eKm6/BFqhLKNuSmWJ+xRHMPyw1QUvSXjVpfK7EK2G6gLw85DWoO1+c8JQPH75tAdhaNZ9w/6r3UtMVisfwbL/EiRNgSE2Y3AAAAAElFTkSuQmCC","orcid":"","institution":"Southern Medical University","correspondingAuthor":true,"prefix":"","firstName":"Jinguo","middleName":"","lastName":"Zhai","suffix":""},{"id":334692726,"identity":"57ec3787-7782-4f11-9590-7e74f9f78acd","order_by":3,"name":"Xueyan Wang","email":"","orcid":"","institution":"Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xueyan","middleName":"","lastName":"Wang","suffix":""},{"id":334692727,"identity":"aa33c10f-1a54-4e86-82d0-6b425b8511aa","order_by":4,"name":"Defang Wu","email":"","orcid":"","institution":"Dongguan Maternal and Child Health Care Hospital","correspondingAuthor":false,"prefix":"","firstName":"Defang","middleName":"","lastName":"Wu","suffix":""},{"id":334692729,"identity":"b06eb3e3-d4d7-4cfa-a2e6-fadd00927475","order_by":5,"name":"Zheng Yao","email":"","orcid":"","institution":"Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zheng","middleName":"","lastName":"Yao","suffix":""}],"badges":[],"createdAt":"2024-07-21 10:53:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4776419/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4776419/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12884-025-07972-8","type":"published","date":"2025-08-18T16:29:45+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":62819295,"identity":"16bb18ea-0b9f-448b-8b43-31fbb95b1507","added_by":"auto","created_at":"2024-08-19 23:52:51","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":118628,"visible":true,"origin":"","legend":"\u003cp\u003eAUC for predicting the probability of cephalic dystocia via the logistic regression model based on LASSO\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure legend: \u003c/strong\u003eModel_1: Model Logistic_53, Model _2: Model Logistic_20. The horizontal axis represents specificity, and the vertical axis represents the sensitivity of risk prediction. AUC: Area under the curve.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-4776419/v1/87a4ab75893f0cc402bbd853.png"},{"id":62819296,"identity":"f6899671-5f10-46cc-86fa-f32c861aacb2","added_by":"auto","created_at":"2024-08-19 23:52:51","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":117630,"visible":true,"origin":"","legend":"\u003cp\u003eModel Logistic_20 variables and coefficients\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure legend: \u003c/strong\u003eThe horizontal axis represents the final variable that enters the model. The vertical axis represents the coefficient of the variable in the regression equation for each variable. A positive coefficient implies a positive correlation, and a negative coefficient implies a negative correlation.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-4776419/v1/c898a8a563d783a711dc238c.png"},{"id":62819298,"identity":"ef7c3adb-e061-4257-9404-62fac9feeffc","added_by":"auto","created_at":"2024-08-19 23:52:51","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":109401,"visible":true,"origin":"","legend":"\u003cp\u003eAUC for predicting the probability of cephalic dystocia via the decision tree method\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure legend: \u003c/strong\u003eModel 1: Model_CART, Model 1 2: Model_C4.5. The horizontal axis indicatesthe specificity of the risk prediction. The vertical axis indicatesthe sensitivity of the risk prediction. AUC: Area under the curve.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-4776419/v1/aa6f7455d0cdc1244ad59e36.png"},{"id":62819299,"identity":"a3799474-d4d1-4c50-a34e-b896693a14b4","added_by":"auto","created_at":"2024-08-19 23:52:51","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":155640,"visible":true,"origin":"","legend":"\u003cp\u003eC4.5 decision tree model\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-4776419/v1/314fcb0575b9d0b8055ad261.png"},{"id":62819998,"identity":"76e0f5d3-7363-43e3-aad7-76a25b48f268","added_by":"auto","created_at":"2024-08-20 00:00:51","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":40755,"visible":true,"origin":"","legend":"\u003cp\u003eThe top 20 variables in terms of the mean decrease in Gini\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-4776419/v1/f4774a4d82092474580e3850.png"},{"id":62820278,"identity":"ac89dc83-dd7a-4458-ab88-83d9f1435d92","added_by":"auto","created_at":"2024-08-20 00:08:52","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1169267,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4776419/v1/67c45e2c-d904-4188-8210-2e03de6ebb80.pdf"},{"id":62819301,"identity":"d4bc21ca-a552-4594-af93-b7ff25b68c82","added_by":"auto","created_at":"2024-08-19 23:52:51","extension":"tiff","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":8368758,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary Fig. 1\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFile format: tiff\u003c/p\u003e\n\u003cp\u003eTitle of figure: AUC for predicting the probability of cephalic dystocia via the random forest method\u003c/p\u003e\n\u003cp\u003eDescription of figure: The horizontal axis indicates the specificity of the risk prediction. The vertical axis indicates the sensitivity of the risk prediction. AUC: Area under the curve.\u003c/p\u003e","description":"","filename":"SupplementaryFig1.tiff","url":"https://assets-eu.researchsquare.com/files/rs-4776419/v1/1557f58cfe9fa7c5b2e247b6.tiff"},{"id":62819302,"identity":"3d5f2f5e-654c-4138-aba8-97a20884c32a","added_by":"auto","created_at":"2024-08-19 23:52:51","extension":"tiff","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":9567480,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary Fig. 2\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFile format: tiff\u003c/p\u003e\n\u003cp\u003eTitle of figure: AUC for predicting the probability of cephalic dystocia via the ML models\u003c/p\u003e\n\u003cp\u003eDescription of figure: The horizontal axis indicates the specificity of the risk prediction. The vertical axis indicates the sensitivity of the risk prediction. AUC: Area under the curve.\u003c/p\u003e","description":"","filename":"SupplementaryFig2.tiff","url":"https://assets-eu.researchsquare.com/files/rs-4776419/v1/8b5559434d5461f00691069e.tiff"}],"financialInterests":"No competing interests reported.","formattedTitle":"Development and validation of a machine learning model for prediction of cephalic dystocia","fulltext":[{"header":"Background","content":"\u003cp\u003eAs of 2021, the World Health Organization (WHO) reported a global cesarean section rate of approximately 21%, with projections suggesting that it will approach 30% by 2030 [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. This rate has been increasing annually. Dystocia is defined as a slow progression of labor, insufficient cervical dilation, and/or lack of fetal head descent [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. It is a common indication for cesarean section. Cephalic dystocia, the most prevalent type of dystocia during childbirth, refers to difficulties in giving birth to the foetus when the head is the first presenting part [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. In some countries, approximately 10% of full-term cephalic pregnancies are complicated during childbirth. Among these cases, 75\u0026thinsp;~\u0026thinsp;80% result in natural childbirth, whereas 20\u0026thinsp;~\u0026thinsp;25% require a cesarean section [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Cephalic dystocia, a multifactorial phenomenon, arises from complex interactions involving fetal head flexion, rotation, and descent [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. It is primarily characterized by slow or prolonged labor progression, posing risks to both mothers and babies while also negatively impacting the birthing experience [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. This situation typically arises after a woman has been in labor for some time, making it difficult to distinguish it from normal cephalic deliveries before birth [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. The timely diagnosis and prediction of cephalic dystocia remain key challenges in obstetrics.\u003c/p\u003e \u003cp\u003eHowever, there are currently a limited number of predictive methods available for cephalic dystocia. Friedman defined the normal labor process in 1954 and introduced the classic \"S\" curve labor graph [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. This graph allows for the identification of prolonged or stalled labor by surpassing specified time limits, aiding clinical professionals in timely intervention. Nevertheless, Friedman\u0026rsquo;s labor graph has faced scrutiny in recent years because it does not fully align with real-world clinical scenarios [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. At present, digital vaginal examination (DVE) is the standard method for assessing labour progress. However, the results depend on the clinician's experience and subjective judgement, with the possibility of interexaminer error[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Additionally, it is an invasive procedure that carries a risk of infection [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. The International Society of Ultrasound in Obstetrics and Gynecology has issued guidelines for the use of ultrasound during childbirth [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Ultrasound imaging has been proven to be more accurate and reproducible than DVE in determining foetal head position and descent into the pelvic region [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Nevertheless, relying solely on a single clinical examination cannot successfully predict the occurrence of head position dystocia. Maternal psychological factors and social factors may also influence labour outcomes. Integrating these factors could contribute to a more comprehensive approach to managing and predicting cephalic dystocia.\u003c/p\u003e \u003cp\u003eMachine learning (ML), which is superior to traditional statistics, can harness extensive data to construct robust predictive models. ML has been widely applied in the fields of obstetrics and gynecology [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. It can assess the risk of disease during pregnancy and predict the mode of childbirth. For example, it can predict the risk of preeclampsia [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], shoulder dystocia [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e], and preterm birth [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn this study, we developed a model using machine learning methods to predict cephalic dystocia. This model is intended to supplement existing approaches and, when integrated into clinical practice, may enhance labor monitoring, assist clinicians in deciding on surgical childbirth, reduce the risk of emergency cesarean sections, and potentially improve maternal and neonatal outcomes.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eThis cross-sectional study was conducted in the Department of Obstetrics and Gynecology at Nanfang Hospital, Southern Medical University, from January 2022 to December 2023. This hospital is a university-based tertiary medical center. Informed consent was obtained from all participants. The study was approved by the Ethics Committee of the Nanfang Hospital of Southern Medical University (No: NFEC-2021-370).\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePatient selection\u003c/h2\u003e \u003cp\u003eThe women who met the inclusion criteria were aged between 18 and 45 years, had a full-term pregnancy, were first-time mothers, carried a single live foetus in a cephalic presentation confirmed by ultrasound, and expressed a desire for a trial of vaginal childbirth. We excluded pregnant women with mental or cognitive disorders, severe pelvic deformities that preclude vaginal childbirth, abnormalities in the birth canal, severe pregnancy complications, or severe internal medical conditions. The data of a total of 320 eligible individuals were collected, of whom 18 had incomplete data. Ultimately, this study included 302 women to establish the predictive model.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eData understanding and collection\u003c/h2\u003e \u003cp\u003eCephalic dystocia refers to obstructed labour occurring in the cephalic position and requiring surgery (cesarean section or vaginal assistance) [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Vaginal-assisted childbirth includes the use of forceps, fetal head aspiration, and manual rotation of the fetal head. Abnormal foetal head position, such as a persistent posterior occipital position, persistent lateral occipital position, and natural childbirth after free-hand rotation of the foetal head, is also a type of cephalic dystocia [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. In certain cases of cephalic childbirth, erroneous judgment can lead to vaginal birth resulting in stillbirth, neonatal death, or intracranial hemorrhage, subsequently causing cerebral palsy or severe intellectual disability. These cases are also classified as cephalic dystocia [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe information of the participating women was stored anonymously. The researchers collected a total of 53 data items, including demographic information (age, education background, occupation, monthly income, etc.) and psychological data. (1) Childbirth Self-Efficacy Inventory (CBSEI-32): This questionnaire was developed by Lowe NK [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] in 1993 and adapted into Chinese by Gao LL et al. [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. It consists of 32 items with a total score of 320. Higher scores indicate stronger self-efficacy in late pregnancy. The Cronbach's α coefficient is 0.96. (2) Pregnancy-Related Anxiety Scale (PrAS): This scale was developed by Brunton RJ et al. [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e] in 2019 and adapted into Chinese by Wu Y [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. It consists of 32 items with a total score of 132, where a score of 75.5 or higher indicates higher anxiety levels. The Cronbach's α coefficient is 0.861. (3) Childbirth Attitude Questionnaire (CAQ): This questionnaire was developed by Lowe NK [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e] in 2000 and adapted into Chinese by Zhang Ming [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. It consists of 16 items with a total score of 64. The Cronbach's α coefficient is 0.882. In addition, the researchers extracted maternal obstetric treatment data from the obstetric electronic medical records system, including the number of pregnancies, fundal height, maternal abdominal circumference (AC), fetal ultrasound parameters, whether oxytocin was administered, whether artificial rupture of membranes (ARM) was performed, whether labor analgesia was used, and other records. Two researchers regularly reconciled the data.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eData processing and statistical analysis\u003c/h2\u003e \u003cp\u003eWe processed the extracted data to generate a database that facilitates algorithm selection. Missing values were imputed using predictive mean matching with the \u0026ldquo;mice\u0026rdquo; package in R version 4.3.2 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.r-project.org/\u003c/span\u003e\u003cspan address=\"https://www.r-project.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), and log\u003csup\u003e2\u003c/sup\u003e transformation was applied to continuous variables with nonnormal distributions. Descriptive statistics were calculated for the baseline characteristics. Continuous variables are expressed as the mean (SD) or median (P25, P75) and were analysed using two-sample t tests or Kruskal‒Wallis tests, depending on normality. Categorical variables are presented as counts or percentages and were analysed using chi-square tests or Fisher's exact tests.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eModel development\u003c/h2\u003e \u003cp\u003eUsing R version 4.3.2, we randomly divided 302 participants into a training set and a validation set at a 75%/25% ratio. The model was constructed based on the training set data. Given the small sample size, we employed 10-fold cross-validation with 5 repeats in each model to enhance the model's credibility. In each fold, 75% of the data were used as the training set, and 25% were used as the test set. The model's performance was evaluated on an independent validation set. The 75%/25% split ratio is commonly used in machine learning algorithms for medium or small sample sizes because it helps build robust models and allows for effective evaluation within a limited sample size [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. We employed three machine learning approaches: LASSO-based logistic regression, decision tree, and random forest methods. In ML algorithms, LASSO regression can help select important features and reduce model complexity. Decision tree methods are easy to understand and interpret, with advantages in handling categorical variables. The random forest algorithm can provide a comprehensive measure of feature importance, offering higher prediction accuracy and helping to reduce the risk of overfitting [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eTwo iterations were conducted for the LASSO-based logistic regression model. The first model, logistic_53, included all 53 variables. The second model, logistic_20, included 20 variables that were identified based on their differences between dystocia and nondystocia women and were ranked among the top 20 in terms of P values in logistic regression analysis. The parameter α was set to 1, and λ was selected from the range 10^seq (2, -2, by -0.1). The optimal λ was determined by maximizing the AUC over 10 repeated cross-validation iterations\u003c/p\u003e \u003cp\u003eThe decision tree generated two models using 22 variables selected based on logistic regression. The first model was the CART classification tree, with complexity parameters ranging from 0.01 to 0.5 in increments of 0.01. The second model was the C4.5 tree, with parameters including boosting iterations set to 10, 20, or 30 and winnowing set to either TRUE or FALSE. The optimal parameter or parameter combination was selected based on the highest AUC after 10 repetitions of cross-validation.\u003c/p\u003e \u003cp\u003eFor the random forest model, 53 variables in the training set were used. The selection of predictor variables was based on minimizing cross-validation error, with the parameter being the number of randomly selected predictors, ranging from 4 to 10. The optimal parameters or parameter combinations were determined by selecting those with the highest AUC after 10 repeated cross-validation iterations. Ultimately, the top 20 variables were selected based on the mean decrease Gini criterion, following the principle of minimizing cross-validation error.\u003c/p\u003e \u003cp\u003eThe model's performance was evaluated using precision, recall, accuracy, p value, and the AUC with a 95% confidence interval (CI). An AUC of 0.5 to \u0026lt;\u0026thinsp;0.7 indicates low model accuracy, 0.7 to 0.9 indicates moderate accuracy, and \u0026gt;\u0026thinsp;0.9 indicates high accuracy [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003ePatient characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBetween January 2022 and December 2023, a total of 302 pregnant women who met the inclusion criteria were selected for the study. Among the study sample, 62 participants (20.5%) experienced cephalic dystocia, while 240 participants (79.5%) had successful natural births. Age, height, body mass index (BMI), educational background, average monthly household income, etc., and demographic characteristicswere not significantly different between the training set and validation set (Table 1).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1\u003c/strong\u003e. Demographic characteristics of the training set (\u003cem\u003en\u003c/em\u003e=227) and testing set (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;75)\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.24050632911393%\"\u003e\n \u003cp\u003e\u003cstrong\u003eCharacteristic\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\"\u003e\n \u003cp\u003e\u003cstrong\u003eTraining set (\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003en\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e=227)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eTesting set (\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003en\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e=75)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.019891500904158%\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eP value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.24050632911393%\" valign=\"top\"\u003e\n \u003cp\u003eAge\u0026nbsp;(year)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" valign=\"top\"\u003e\n \u003cp\u003e28.00 (3.68)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e28.09\u0026nbsp;(3.48)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.019891500904158%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.84\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.24050632911393%\" valign=\"top\"\u003e\n \u003cp\u003ePregnancy days at admission\u0026nbsp;(day)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" valign=\"top\"\u003e\n \u003cp\u003e275.65\u0026nbsp;(6.37)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e275.03\u0026nbsp;(7.08)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.019891500904158%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.47\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.24050632911393%\" valign=\"top\"\u003e\n \u003cp\u003eMaternal height\u0026nbsp;(cm)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" valign=\"top\"\u003e\n \u003cp\u003e159.58\u0026nbsp;(4.82)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e158.89\u0026nbsp;(5.19)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.019891500904158%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.24050632911393%\" valign=\"top\"\u003e\n \u003cp\u003eMaternal weight before pregnancy\u0026nbsp;(kg)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" valign=\"top\"\u003e\n \u003cp\u003e50.37\u0026nbsp;(7.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e51.02\u0026nbsp;(6.55)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.019891500904158%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.24050632911393%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 1.\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cem\u003e(continued)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.019891500904158%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.24050632911393%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eCharacteristic\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eTraining set (\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003en\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e=227)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eTesting set (\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003en\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e=75)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.019891500904158%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eP value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.24050632911393%\" valign=\"top\"\u003e\n \u003cp\u003eMaternal BMI before pregnancy\u0026nbsp;(kg/m\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" valign=\"top\"\u003e\n \u003cp\u003e19.76\u0026nbsp;(2.59)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.869801084990957%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e20.19\u0026nbsp;(2.29)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.019891500904158%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.20\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eCurrent weight\u0026nbsp;(kg)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e64.26\u0026nbsp;(7.78)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e65.65\u0026nbsp;(8.85)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e0.19\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eWeight gain during pregnancy\u0026nbsp;(kg)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e13.88\u0026nbsp;(3.87)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e14.62\u0026nbsp;(4.90)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e0.38\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eCurrent BMI\u0026nbsp;(kg/m\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e25.21\u0026nbsp;(2.66)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e25.98\u0026nbsp;(3.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eMedical payment method\u003cem\u003e\u0026nbsp;\u003c/em\u003e(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e0.15\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eUrban employee basic medical insurance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e180\u0026nbsp;(79.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e57\u0026nbsp;(76.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eUrban resident basic medical insurance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e3\u0026nbsp;(1.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e1\u0026nbsp;(1.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003ecompletely publicly funded\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e1\u0026nbsp;(0.40)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e3\u0026nbsp;(4.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003efully self-funded\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e43\u0026nbsp;(18.90)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e14\u0026nbsp;(18.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eFamily per capita monthly income\u0026nbsp;(CNY)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e0.72\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003e＜5000\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e16\u0026nbsp;(7.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e4\u0026nbsp;(5.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003e5000～9999\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e91\u0026nbsp;(40.10)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e26\u0026nbsp;(34.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003e10000～14999\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e63\u0026nbsp;(27.80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e22\u0026nbsp;(29.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026ge;15000\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e57\u0026nbsp;(25.10)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e23\u0026nbsp;(30.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eMaternal educational background\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e0.47\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eJunior high school and below\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e42\u0026nbsp;(18.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e10\u0026nbsp;(13.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eHigh school and technical secondary school\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e26\u0026nbsp;(11.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e14\u0026nbsp;(18.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eCollege and undergraduate\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e150\u0026nbsp;(66.10)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e43\u0026nbsp;(57.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eBachelor degree or above\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e9\u0026nbsp;(4.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e8\u0026nbsp;(10.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eWhether this pregnancy is planned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e0.97\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eNo\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e67\u0026nbsp;(29.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e22\u0026nbsp;(29.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eYes\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e160\u0026nbsp;(70.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e53\u0026nbsp;(70.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003ePregnancy mode\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e0.08\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eNature conceived\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e211(93.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e74\u0026nbsp;(98.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eART\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e16(7.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e1(1.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eWhether continue to work in the third trimester of pregnancy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e0.77\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eNo\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e86(37.90)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e27\u0026nbsp;(36.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003eYes\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e141(62.10)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e48\u0026nbsp;(64.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003ePregnancy times, \u003cem\u003en\u003c/em\u003e (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e0.51\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003e1\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e164\u0026nbsp;(72.20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e59\u0026nbsp;(78.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003e2\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e50\u0026nbsp;(22.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e15\u0026nbsp;(20.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003e3\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e11\u0026nbsp;(4.80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e1\u0026nbsp;(1.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"39.16967509025271%\" valign=\"top\"\u003e\n \u003cp\u003e4\u0026nbsp;(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.63176895306859%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e2\u0026nbsp;(0.90)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.270758122743683%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0\u0026nbsp;(0.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.927797833935019%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eContinuous variables are expressed in mean\u0026plusmn;standard deviation (SD) or median (25th\u0026ndash;75th percentiles). Categorical variables were expressed as frequencies (percentages). Abbreviations: BMI: body mass index, CNY: China Yuan, ART: assisted reproductive technology.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLASSO-based\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003elogistic regression model d\u003c/strong\u003e\u003cstrong\u003eescription\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor the training dataset, Model Logistic_53 exhibited comparatively lower performance across various performance metrics compared to Model Logistic_20. In the testing dataset, Logistic_53 exhibited higher recall (0.967 vs 0.933) than logistic_20. However, Logistic_20 outperformed Logistic_53 in terms of precision, accuracy and AUC, as shown in Table 2. The validation performance, shown in Fig. 1, indicates a significant enhancement in prediction performance for Logistic_20.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2.\u003c/strong\u003e Performance parameters of the five machine learning prediction models in the training and testing sets\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"15.463917525773196%\"\u003e\n \u003cp\u003e\u003cstrong\u003eData set\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"21.649484536082475%\"\u003e\n \u003cp\u003e\u003cstrong\u003ePredictive models\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.34020618556701%\"\u003e\n \u003cp\u003e\u003cstrong\u003ePrecision\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.278350515463918%\"\u003e\n \u003cp\u003e\u003cstrong\u003eRecall\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.278350515463918%\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eAccuracy\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.34020618556701%\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003evalue\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"21.649484536082475%\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC (95%\u003cem\u003e\u0026nbsp;CI\u003c/em\u003e)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"15.463917525773196%\" rowspan=\"5\" valign=\"top\"\u003e\n \u003cp\u003eTraining set\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"21.649484536082475%\" valign=\"top\"\u003e\n \u003cp\u003eLogistic_20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.34020618556701%\" valign=\"top\"\u003e\n \u003cp\u003e0.890\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.278350515463918%\" valign=\"top\"\u003e\n \u003cp\u003e0.989\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.278350515463918%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.894\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.34020618556701%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e1.05e-04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"21.649484536082475%\" valign=\"top\"\u003e\n \u003cp\u003e0.760 (0.688,0.833)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003eLogistic_53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e0.864\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" valign=\"top\"\u003e\n \u003cp\u003e0.989\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.868\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e5.01e-06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003e0.697 (0.625,0.768)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003eRandom forest\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" valign=\"top\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eNaN\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003e1.000 (1.000,1.000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003eDecision tree CART\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e0.855\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" valign=\"top\"\u003e\n \u003cp\u003e0.983\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.855\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e6.01e-06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003e0.673 (0.602,0.743)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003eDecision tree C4.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e0.928\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" valign=\"top\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.938\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e5.12e-04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003e0.851 (0.785,0.917)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"15.463917525773196%\" rowspan=\"5\" valign=\"top\"\u003e\n \u003cp\u003eTesting set\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"21.649484536082475%\" valign=\"top\"\u003e\n \u003cp\u003eLogistic_20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.34020618556701%\" valign=\"top\"\u003e\n \u003cp\u003e0.933\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.278350515463918%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.933\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.278350515463918%\" valign=\"top\"\u003e\n \u003cp\u003e0.893\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"11.34020618556701%\" valign=\"top\"\u003e\n \u003cp\u003e1.00e+00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"21.649484536082475%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.833 (0.713,0.953)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003eLogistic_53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e0.892\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.967\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" valign=\"top\"\u003e\n \u003cp\u003e0.880\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e1.82e-01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.609756097560975%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.750 (0.617,0.883)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003eRandom forest\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e0.892\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.967\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" valign=\"top\"\u003e\n \u003cp\u003e0.880\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e1.82e-01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.609756097560975%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.750 (0.617,0.883)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003eDecision tree CART\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e0.919\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.950\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" valign=\"top\"\u003e\n \u003cp\u003e0.893\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e7.24e-01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.609756097560975%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.808 (0.682,0.935)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.609756097560975%\" valign=\"top\"\u003e\n \u003cp\u003eDecision tree C4.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e0.919\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.950\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.975609756097562%\" valign=\"top\"\u003e\n \u003cp\u003e0.893\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.414634146341463%\" valign=\"top\"\u003e\n \u003cp\u003e7.24e-01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.609756097560975%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e0.767 (0.634,0.899)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003eConsequently, we selected Model Logistic_20 as our preferred predictive model because of its superior overall performance. The final model encompasses 18 variables, including the internodal diameter of maternal ischial bone, fetal biparietal diameter (BPD), fetal humeral diameter (HL), fetal biophysical profile score (BPP), artificial rupture of membranes, doula, and other variables (Fig. 2). The final regression equation was as follows: logit P= －0.18－0.038 \u0026times; CBSEI-32 score－0.36 \u0026times; Fetal biophysical profile score＋0.867 \u0026times; Fetal biparietal diameter－1.38 \u0026times; Fetal humeral diameter＋0.014 \u0026times; Husband\u0026rsquo;s height－0.523 \u0026times; Internodal diameter of maternal ischial bone－0.141 \u0026times; Maternal interiliac spine diameter＋0.046 \u0026times; Medical payment method＋0.058 \u0026times; Pregnancy days at admission＋0.672 \u0026times; Pregnancy mode－0.153 \u0026times; Pregnancy times＋1.034 \u0026times; Whether artificial rupture of membranes－0.122 \u0026times; Whether continue to work in the third trimester of pregnancy－0.473\u0026times; Whether there is accompanied labor＋0.143 \u0026times; Whether induced labor with water sac＋ 0.905 \u0026times; Whether there is Doula－0.224 \u0026times; Whether there is labor analgesia－0.137\u0026times; Whether this pregnancy is planned.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDecision\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003etree\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;model description\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAccording to logistic regression, 22 variables were selected to construct the\u0026nbsp;decision tree\u0026nbsp;model. In the training dataset, Model_C4.5 outperformed Model_CART across various indicators, as\u0026nbsp;described\u0026nbsp;in detail in Table 2. After cross-validation, in the testing dataset, Model_CART and Model C4.5 exhibited identical precision, recall, and accuracy, with a\u0026nbsp;greater\u0026nbsp;AUC for Model_C4.5\u0026nbsp;than for\u0026nbsp;Model CART (0.808 vs 0.767), as shown in Fig.\u0026nbsp;3.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eFive variables played a crucial role in the construction of the decision tree: CBSEI_32 score, BPP, ARM, doula, and HL. The resulting decision tree is depicted in Fig. 4.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRandom\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eforest model description\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFollowing the principle of minimizing cross-validation error, we selected the top 20 variables based on the mean decrease Gini coefficient , where a larger \u0026quot;mean decrease Gini\u0026quot; value indicates a greater contribution of the feature to improving classification performance. Among the crucial features were the CBSEI_32 score, fetal head circumference (HC), maternal height, AC, the S/D value of umbilical blood flow, HC, and other variables, as illustrated in Fig. 5. A classification model was constructed on the basis of these selected top 20 variables, with the optimal parameter set to 8. The model achieved an AUC of 1.000 (95% CI: 1.000 to 1.000) on the training set and 0.750 (95% CI: 0.617 to 0.883) on the testing set, as shown in Supplementary Fig. 1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOverall performance comparison\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eModel logistic_20 appears to be a promising model, as it performs well across multiple performance metrics, including precision, recall, accuracy, and AUC values, as well as small P values and relatively narrow confidence intervals. We generated AUC curves for all the models to visualize these results, as shown in Supplementary Fig. 2.\u0026nbsp;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIdentifying cephalic dystocia prior to labor is highly challenging because of multiple influencing factors. Previous studies have reported relevant risk factors but have seldom developed predictive models [29-32]. This study utilized three machine learning methods to develop and validate a predictive model for cephalic dystocia.\u0026nbsp;Fetal\u0026nbsp;ultrasound parameters, maternal anthropometric measurements, psychological factors, and obstetric records\u0026nbsp;were collected. The model\u0026nbsp;can assist\u0026nbsp;clinicians in evaluating the probability of cephalic dystocia in primiparous women at term, facilitating the early identification of high-risk\u0026nbsp;women. Additionally, LASSO regression highlighted several crucial variables deserving further investigation.\u003c/p\u003e\n\u003cp\u003eUltrasound plays a crucial role in assessing fetal size and position, assisting clinicians in predicting the risk of dystocia [7]. In our study, we found that BPD, HL, and BPP\u0026nbsp;were\u0026nbsp;significant in the predictive model.\u0026nbsp;The\u0026nbsp;BPD can be used to estimate the likelihood of\u0026nbsp;a foetus\u0026nbsp;successfully passing through the birth canal. Shinohara S et al. [29]\u0026nbsp;reported\u0026nbsp;that a\u0026nbsp;greater\u0026nbsp;BPD might increase the risk of dystocia. In this study, a positive correlation between BPD and dystocia was also observed. A longitudinal cohort study of 2802 pregnant women\u0026nbsp;revealed\u0026nbsp;that fetuses of obese women had significantly longer\u0026nbsp;HLs, increasing the risk of dystocia [33]. In our study,\u0026nbsp;HL\u0026nbsp;was negatively correlated\u0026nbsp;with cephalic dystocia. A larger HL may indicate more mature fetal skeletal development, potentially aiding in withstanding pressure and facilitating passage through the birth canal during childbirth. However, excessively\u0026nbsp;high\u0026nbsp;HL is typically associated with macrosomia\u0026nbsp;[33].\u0026nbsp;According to\u0026nbsp;a multinational longitudinal study by the\u0026nbsp;WHO, the HL of a\u0026nbsp;foetus\u0026nbsp;at 37–40 weeks is generally between\u0026nbsp;67 mm\u0026nbsp;and\u0026nbsp;69 mm [34]. An HL within this normal range may be more conducive to smooth delivery.\u003c/p\u003e\n\u003cp\u003eThe\u0026nbsp;BPP assessment helps clinicians determine whether the fetus is healthy or at risk of intrauterine hypoxia [35]. A higher BPP indicates better fetal health and is associated with a\u0026nbsp;greater\u0026nbsp;success rate of vaginal childbirth [36]. Similar to the negative correlation found in this study between\u0026nbsp;foetal\u0026nbsp;physiological score and difficult cephalic presentation, the BPP\u0026nbsp;has\u0026nbsp;predictive value. Furthermore, in our study,\u0026nbsp;the internodal diameter of maternal ischial bone\u0026nbsp;and the interiliac spine diameter\u0026nbsp;were\u0026nbsp;negatively correlated with cephalic dystocia. A larger\u0026nbsp;internodal diameter of maternal ischial bone\u0026nbsp;and interiliac spine diameter\u0026nbsp;in parturients signifies a wider pelvis, which can facilitate better fetal head engagement and provide more space in the birth canal..This reduces birth canal resistance, lowering the risk of\u0026nbsp;foetal\u0026nbsp;head obstruction or impaction and decreasing the risk of cephalic dystocia [30]. By measuring\u0026nbsp;the internodal diameter of maternal ischial bone\u0026nbsp;and the interiliac spine diameter, healthcare providers can devise more personalized childbirth plans. Women with wider pelvises may be encouraged to pursue natural childbirth, while those with narrower pelvises may require early\u0026nbsp;intervention,\u0026nbsp;such as planned cesarean section,\u0026nbsp;to ensure maternal and infant safety.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eARM is typically performed after normal uterine contractions have been induced using oxytocin or mechanical methods. [37]. In our study,\u0026nbsp;an\u0026nbsp;ARM was identified as a risk factor for cephalic dystocia. ARM alone may impede cervical ripening, delay the labor process, and increase the risk of chorioamnionitis and umbilical cord prolapse [37].\u0026nbsp;Research has shown that using an ARM does not shorten the first or second stages of labor and tends to increase the risk of cesarean section\u0026nbsp;[38]. The American College of Obstetricians and Gynecologists also\u0026nbsp;advised\u0026nbsp;against routine ARM solely to prevent labor prolongation [39]. Furthermore, labor analgesia\u0026nbsp;was\u0026nbsp;negatively correlated with cephalic dystocia in our study. Labor analgesia can\u0026nbsp;reduce\u0026nbsp;pain stress responses\u0026nbsp;and\u0026nbsp;aid\u0026nbsp;uterine contractions.\u0026nbsp;Moreover,\u0026nbsp;it enhances the willingness of women in labor to undergo vaginal childbirth, enabling better cooperation with medical guidance and promoting a smoother labor process [40]. While labor analgesia may prolong the duration of labor, research suggests that this prolongation does not\u0026nbsp;increase\u0026nbsp;the incidence of postpartum hemorrhage or fetal distress [41].\u0026nbsp;Thus, healthcare providers are strongly encouraged to offer labor analgesia to eligible women, thereby decreasing the risk of cephalic dystocia.\u003c/p\u003e\n\u003cp\u003eIn this study, water sac induction was considered a risk factor for cephalic dystocia. Water sac induction primarily uses Foley catheters and COOK cervical ripening balloons. Obstetricians typically place\u0026nbsp;balloons\u0026nbsp;in the cervical canal to achieve mechanical dilation. By exerting pressure on the cervix, the balloon stimulates the release of endogenous prostaglandins, facilitating cervical ripening [42]. However, research shows that pregnant women\u0026nbsp;who undergo\u0026nbsp;Foley catheter balloon placement have\u0026nbsp;greater\u0026nbsp;infection rates (OR=1.50, 95% CI: 1.07 to 2.09)\u0026nbsp;than\u0026nbsp;those\u0026nbsp;who undergo\u0026nbsp;pharmacological induction methods, which increases the risk of dystocia [43]. Additionally, the duration and volume of catheter balloon placement may also influence the progression of labor, necessitating further exploration in future studies.\u003c/p\u003e\n\u003cp\u003eThis study also\u0026nbsp;revealed\u0026nbsp;that higher scores on the childbirth self-efficacy scale are associated with a lower likelihood of cephalic dystocia. Childbirth self-efficacy reflects women's expectations of childbirth outcomes and confidence in their abilities [20]. Research has shown a strong correlation between childbirth self-efficacy and childbirth fear [32]. Consequently, women with lower self-efficacy may be more susceptible to experiencing tension and fear, triggering catecholamine release that inhibits effective uterine contractions, thereby hindering labor progress [44]. Healthcare providers should actively engage in prenatal education and provide psychological support to pregnant women.\u0026nbsp;The\u0026nbsp;presence of a doula is also a significant predictor\u0026nbsp;of cephalic dystocia. Experienced doulas can monitor labor progress, provide childbirth assistance, enhance maternal and infant safety, promptly identify dystocia risks, and simultaneously increase maternal childbirth self-efficacy through psychological support [45]. In this study, the presence of doula was positively correlated with dystocia. Doulas may reduce the incidence of dystocia. However, pregnant women who perceive themselves at risk for dystocia may be more likely to seek or be advised to seek doula assistance. As a result, endogeneity could affect the model, leading to estimated coefficients that may not accurately reflect the true relationship. Additionally, our sample size may be insufficient, highlighting the need for further research to explore this issue.\u003c/p\u003e\n\u003cp\u003eThe\u0026nbsp;WHO recommends that women be supported throughout childbirth by selected peers, including midwives, doctors, doulas, spouses or other family members, and friends [46]. In this study, labor support was negatively correlated with cephalic dystocia.\u0026nbsp;The maternal\u0026nbsp;paternal\u0026nbsp;partner\u0026nbsp;is a female partner or family member. Having a companion during labor eliminates the sense of isolation, improves the childbirth experience, and contributes to a smoother birth process [47]. Research indicates that having a spouse as a companion promotes the transition of the husband's role and enhances the marital relationship [48]. However, accompanying childbirth may also have adverse effects [47]. If the companion is not well\u0026nbsp;prepared, they may feel anxious when witnessing the parturient's labor. When the parturient sees negative facial expressions or body language from her companion,\u0026nbsp;this\u0026nbsp;may amplify her negative assessment of the situation, leading to increased anxiety and difficulty in coping with labor pain. Therefore, companions need adequate preparation and training to ensure\u0026nbsp;that\u0026nbsp;they can provide the best support.\u003c/p\u003e\n\u003cp\u003eIn this study, the method of conception\u0026nbsp;was\u0026nbsp;associated with cephalic dystocia. Women who conceive through\u0026nbsp;assisted reproductive technology\u0026nbsp;(ART)\u0026nbsp;are more prone to cephalic dystocia. These women typically exhibit physiological characteristics related to infertility or comorbidities.\u0026nbsp;These individuals\u0026nbsp;are at\u0026nbsp;increased\u0026nbsp;risk of developing gestational diabetes, gestational hypertension, and abnormal amniotic fluid levels, which can affect fetal intrauterine growth and development [49]. Additionally, they may experience a heavier psychological burden during pregnancy and childbirth due to long-term ovulation induction or early pregnancy maintenance drugs [50]. As a result,\u0026nbsp;these women\u0026nbsp;are at\u0026nbsp;greater\u0026nbsp;risk of cephalic dystocia and may require emergency cesarean sections or assisted deliveries. Studies also show that the cesarean section rate is\u0026nbsp;greater for\u0026nbsp;women who conceive through ART\u0026nbsp;than for\u0026nbsp;those who conceive naturally [51]. Clinical doctors and\u0026nbsp;midwives\u0026nbsp;should pay special attention to the childbirth process of women who become pregnant through ART. Conversely, women who achieve pregnancy through natural conception appear to have a lower risk of cephalic dystocia in this study. This may be due to their preparation before conception, during pregnancy, and\u0026nbsp;during\u0026nbsp;childbirth, as well as their more\u0026nbsp;favourable\u0026nbsp;physiological and psychological conditions for a smooth birth.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStrengths and limitations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur study has several strengths. We collected a comprehensive set of risk factor data, including foetal ultrasound parameters, maternal anthropometric measurements, mothers' psychological status, and obstetric treatment data. Our research data do not add additional costs for women. By using the LASSO regression algorithm, we removed variables that did not contribute significantly to the prediction of difficult labor, resulting in a more streamlined model with enhanced predictive performance. However, there are several limitations to our study. First, the sample size was not large enough, and external validation with a larger dataset is needed to confirm the effectiveness of the model. Additionally, potential endogeneity among independent variables may have arisen due to interactions between them, necessitating further adjustment and variable selection in a larger sample.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThis study employed ML to develop a predictive model for cephalic dystocia, which demonstrated strong predictive performance. As a supplement to existing methods, the model can assist healthcare providers in the early identification of pregnant women at risk for cephalic dystocia and in making decisions about surgical delivery. Additionally, healthcare providers can focus on modifiable variables during labor, such as labor analgesia, balloon induction, labor companionship, and psychological support, to reduce the risk of dystocia. To further validate the model\u0026apos;s results and enhance its performance, future studies should involve multicenter, large-sample prospective trials.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eLASSO:\u0026nbsp;least absolute shrinkage and selection operator\u003c/p\u003e\n\u003cp\u003eROC: Receiver operating characteristic\u003c/p\u003e\n\u003cp\u003eAUC: area under the ROC curve\u003c/p\u003e\n\u003cp\u003eWHO: World Health Organization\u003c/p\u003e\n\u003cp\u003eDVE: digital vaginal examination\u003c/p\u003e\n\u003cp\u003eML: machine learning\u003c/p\u003e\n\u003cp\u003eCBSEI-32: Childbirth Self-Efficacy Inventory\u003c/p\u003e\n\u003cp\u003ePrAS: Pregnancy-Related Anxiety Scale\u003c/p\u003e\n\u003cp\u003eCAQ: Childbirth Attitude Questionnaire\u003c/p\u003e\n\u003cp\u003eAC: abdominal circumference\u003c/p\u003e\n\u003cp\u003eARM: artificial rupture of membranes\u003c/p\u003e\n\u003cp\u003eCI: confidence interval\u003c/p\u003e\n\u003cp\u003eBMI: body mass index\u003c/p\u003e\n\u003cp\u003eCNY: China Yuan\u003c/p\u003e\n\u003cp\u003eBPD: biparietal diameter\u003c/p\u003e\n\u003cp\u003eHL: humeral diameter\u003c/p\u003e\n\u003cp\u003eBPP: fetal biophysical profile score\u003c/p\u003e\n\u003cp\u003eHC: head circumference\u003c/p\u003e\n\u003cp\u003eART: assisted reproductive technology\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by\u0026nbsp;the\u0026nbsp;Nanfang Hospital of Southern Medical University Ethics Committee (NFEC-2021-370).\u0026nbsp;The methodology adhered to the tenets of the Declaration of Helsinki.\u0026nbsp;All participants signed informed consent forms.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets analysed during the current study are available from the corresponding author upon \u0026nbsp; reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by\u0026nbsp;the\u0026nbsp;Special Higher Education Program of Guangdong Provincial Education Science Planning in 2021(2021GXJK163) and\u0026nbsp;the\u0026nbsp;Quality Engineering Construction Project of Southern Medical University in 2021-Teaching and Research Section of Midwifery (202129). The funders had no role in\u0026nbsp;the\u0026nbsp;study design, data collection and analysis, decision to publish, or preparation of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u003c/strong\u003e\u003cstrong\u003e\u0026rsquo;\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eYMH and XRR should be considered joint first author. YMH: Methodology, investigation, writing - original draft; XRR: Investigation, software, formal analysis, writing-original draft; JGZ: Conceptualization, project administration, supervision, writing - review \u0026amp; editing; XYW: Data curation, formal analysis; DFW and ZY: Investigation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBetran AP, Ye J, Moller AB, Souza JP, Zhang J. Trends and projections of caesarean section rates: global and regional estimates. \u003cem\u003eBMJ Glob Health \u003c/em\u003e2021, 6(6):e005671.\u003c/li\u003e\n\u003cli\u003eRessel GW. ACOG releases report on dystocia and augmentation of labor. \u003cem\u003eAm Fam Physician \u003c/em\u003e2004, 69(5):1290, 1292.\u003c/li\u003e\n\u003cli\u003eCunningham FG, Leveno KJ, Bloom SL, Dashe JS, Hoffman BL, Casey BM, Spong CY. Abnormal Labor. In\u003cem\u003e.\u003c/em\u003e; 2018.\u003c/li\u003e\n\u003cli\u003eCaughey AB, Cahill AG, Guise JM, Rouse DJ. Safe prevention of the primary cesarean delivery. \u003cem\u003eAmerican journal of obstetrics and gynecology \u003c/em\u003e2014, 210(3):179-193.\u003c/li\u003e\n\u003cli\u003eDall\u0026rsquo;Asta A, Ferretti A, Minopoli M, Ghi T. The role of the occiput-spine angle in prolonged labour and delivery outcome. \u003cem\u003eIntrapartum ultrasonography for labor management: labor, delivery and puerperium \u003c/em\u003e2021:319-323.\u003c/li\u003e\n\u003cli\u003eWalker KF, Kibuka M, Thornton JG, Jones NW. Maternal position in the second stage of labour for women with epidural anaesthesia. \u003cem\u003eThe Cochrane database of systematic reviews \u003c/em\u003e2018, 11(11):CD008070.\u003c/li\u003e\n\u003cli\u003eDall\u0026apos;Asta A, Rizzo G, Masturzo B, Di Pasquo E, Schera GBL, Morganelli G, Ramirez ZR, Maqina P, Mappa I, Parpinel G\u003cem\u003e et al\u003c/em\u003e. Intrapartum sonographic assessment of the fetal head flexion in protracted active phase of labor and association with labor outcome: a multicenter, prospective study. \u003cem\u003eAmerican journal of obstetrics and gynecology \u003c/em\u003e2021, 225(2):171.e171-171.e112.\u003c/li\u003e\n\u003cli\u003eFriedman E. The graphic analysis of labor. \u003cem\u003eAmerican journal of obstetrics and gynecology \u003c/em\u003e1954, 68(6):1568-1575.\u003c/li\u003e\n\u003cli\u003eZhang J, Landy HJ, Ware BD, Burkman R, Haberman S, Gregory KD, Hatjis CG, Ramirez MM, Bailit JL, Gonzalez-Quintero VH\u003cem\u003e et al\u003c/em\u003e. Contemporary patterns of spontaneous labor with normal neonatal outcomes. \u003cem\u003eObstetrics and gynecology \u003c/em\u003e2010, 116(6):1281-1287.\u003c/li\u003e\n\u003cli\u003eEnomoto N, Maki S, Nii M, Yamaguchi M, Tamaishi Y, Takakura S, Magawa S, Tanaka K, Tanaka H, Kondo E\u003cem\u003e et al\u003c/em\u003e. Accurate evaluation of the progress of delivery with transperineal ultrasound may improve vaginal delivery: a single-center retrospective study. \u003cem\u003eScientific reports \u003c/em\u003e2023, 13(1):20945.\u003c/li\u003e\n\u003cli\u003eAfework B, Birgoda GT, Leul H, Girma L, Bekelcho T. Frequency of digital vaginal examination and associated factors among mothers admitted to labor and delivery wards in public health facilities of Arba Minch town, southern Ethiopia. \u003cem\u003eInt J Gynecol Obstet \u003c/em\u003e2024, 164(3):1028-1035.\u003c/li\u003e\n\u003cli\u003eGhi T, Eggeb\u0026oslash; T, Lees C, Kalache K, Rozenberg P, Youssef A, Salomon LJ, Tutschek B. ISUOG Practice Guidelines: intrapartum ultrasound. \u003cem\u003eUltrasound in obstetrics \u0026amp; gynecology : the official journal of the International Society of Ultrasound in Obstetrics and Gynecology \u003c/em\u003e2018, 52(1):128-139.\u003c/li\u003e\n\u003cli\u003eSherer DM, Miodovnik M, Bradley KS, Langer O. Intrapartum fetal head position I: comparison between transvaginal digital examination and transabdominal ultrasound assessment during the active stage of labor. \u003cem\u003eUltrasound in obstetrics \u0026amp; gynecology : the official journal of the International Society of Ultrasound in Obstetrics and Gynecology \u003c/em\u003e2002, 19(3):258-263.\u003c/li\u003e\n\u003cli\u003eIslam MN, Mustafina SN, Mahmud T, Khan NI. Machine learning to predict pregnancy outcomes: a systematic review, synthesizing framework and future research agenda. \u003cem\u003eBMC pregnancy and childbirth \u003c/em\u003e2022, 22(1):348.\u003c/li\u003e\n\u003cli\u003eMarić I, Tsur A, Aghaeepour N, Montanari A, Stevenson DK, Shaw GM, Winn VD. Early prediction of preeclampsia via machine learning. \u003cem\u003eAmerican journal of obstetrics \u0026amp; gynecology MFM \u003c/em\u003e2020, 2(2):100100.\u003c/li\u003e\n\u003cli\u003eTsur A, Batsry L, Toussia-Cohen S, Rosenstein MG, Barak O, Brezinov Y, Yoeli-Ullman R, Sivan E, Sirota M, Druzin ML\u003cem\u003e et al\u003c/em\u003e. Development and validation of a machine-learning model for prediction of shoulder dystocia. \u003cem\u003eUltrasound in obstetrics \u0026amp; gynecology : the official journal of the International Society of Ultrasound in Obstetrics and Gynecology \u003c/em\u003e2020, 56(4):588-596.\u003c/li\u003e\n\u003cli\u003eGao C, Osmundson S, Velez Edwards DR, Jackson GP, Malin BA, Chen Y. Deep learning predicts extreme preterm birth from electronic health records. \u003cem\u003eJournal of biomedical informatics \u003c/em\u003e2019, 100:103334.\u003c/li\u003e\n\u003cli\u003eLing LD. Pay attention to cephalic dystocia (article in Chinese). \u003cem\u003eChinese Journal of Practical Gynecology and Obstetrics \u003c/em\u003e1994, 10(04):194-195.\u003c/li\u003e\n\u003cli\u003eQi HB. Identification and treatment of cephalic dystocia (article in Chinese). \u003cem\u003eChinese Journal of Practical Gynecology and Obstetrics \u003c/em\u003e2005, 21(05):264-267.\u003c/li\u003e\n\u003cli\u003eLowe NK. Maternal confidence for labor: development of the Childbirth Self-Efficacy Inventory. \u003cem\u003eResearch in nursing \u0026amp; health \u003c/em\u003e1993, 16(2):141-149.\u003c/li\u003e\n\u003cli\u003eGao LL, Ip WY, Sun K. Validation of the short form of the Chinese Childbirth Self-Efficacy Inventory in Mainland China. \u003cem\u003eResearch in nursing \u0026amp; health \u003c/em\u003e2011, 34(1):49-59.\u003c/li\u003e\n\u003cli\u003eBrunton RJ, Dryer R, Saliba A, Kohlhoff J. The initial development of the Pregnancy-related Anxiety Scale. \u003cem\u003eWomen and birth : journal of the Australian College of Midwives \u003c/em\u003e2019, 32(1):e118-e130.\u003c/li\u003e\n\u003cli\u003eWu Y. Chinese version of Pregnancy-related Anxiety Scale and its reliability and validity test (article in Chinese). \u003cem\u003eMaster.\u003c/em\u003e Kunming Medical University; 2021.\u003c/li\u003e\n\u003cli\u003eLowe NK. Self-efficacy for labor and childbirth fears in nulliparous pregnant women. \u003cem\u003eJournal of psychosomatic obstetrics and gynaecology \u003c/em\u003e2000, 21(4):219-224.\u003c/li\u003e\n\u003cli\u003eZhang M. The research of Antenatal Fear of Childbirth and Its Associated Factors in Pregnant Women (article in Chinese). \u003cem\u003eMaster.\u003c/em\u003e Jilin University; 2016.\u003c/li\u003e\n\u003cli\u003eJoseph VR. Optimal ratio for data splitting. 2022, 15(4):531-538.\u003c/li\u003e\n\u003cli\u003eNgiam KY, Khor IW. Big data and machine learning algorithms for health-care delivery. \u003cem\u003eLancet Oncol \u003c/em\u003e2019, 20(5):e262-e273.\u003c/li\u003e\n\u003cli\u003eJayawant NM. Receiver Operating Characteristic Curve in Diagnostic Test Assessment. \u003cem\u003eJournal of Thoracic Oncology \u003c/em\u003e2010, 5(9):1315-1316.\u003c/li\u003e\n\u003cli\u003eShinohara S, Amemiya A, Takizawa M, Suzuki K. Fetal biparietal diameter as a potential risk factor for prolonged second stage of labor: A retrospective observational cohort study. \u003cem\u003ePLoS One \u003c/em\u003e2020, 15(10):e0239045.\u003c/li\u003e\n\u003cli\u003eJochumsen S, Hegaard HK, Rode L, J\u0026oslash;rgensen KJ, Nathan NO. Maternal factors associated with labor dystocia in low-risk nulliparous women. A systematic review and meta-analysis. \u003cem\u003eSexual \u0026amp; reproductive healthcare : official journal of the Swedish Association of Midwives \u003c/em\u003e2023, 36:100855.\u003c/li\u003e\n\u003cli\u003eOduro AR, Anyorikeya M, Ansah P, Oladokun S, Tei EM, Oduro-Ayeh R, Welaga P, Deh S. Birth preparedness and complications readiness among women in disadvantaged rural districts of Ghana. \u003cem\u003eBMC pregnancy and childbirth \u003c/em\u003e2023, 23(1):728.\u003c/li\u003e\n\u003cli\u003eDwiarini M, Chou HF, Gau ML, Liu CY. Relationship between fear of childbirth, self-efficacy, and length of labor among nulliparous women in Indonesia. \u003cem\u003eMidwifery \u003c/em\u003e2022, 105:103203.\u003c/li\u003e\n\u003cli\u003eZhang C, Hediger ML, Albert PS, Grewal J, Sciscione A, Grobman WA, Wing DA, Newman RB, Wapner R, D\u0026apos;Alton ME\u003cem\u003e et al\u003c/em\u003e\u003cem\u003e.\u003c/em\u003e Association of Maternal Obesity With Longitudinal Ultrasonographic Measures of Fetal Growth: Findings From the NICHD Fetal Growth Studies-Singletons. \u003cem\u003eJAMA Pediatr \u003c/em\u003e2018, 172(1):24-31.\u003c/li\u003e\n\u003cli\u003eKiserud T, Piaggio G, Carroli G, Widmer M, Carvalho J, Neerup Jensen L, Giordano D, Cecatti JG, Abdel Aleem H, Talegawkar SA\u003cem\u003e et al\u003c/em\u003e\u003cem\u003e.\u003c/em\u003e The World Health Organization Fetal Growth Charts: A Multinational Longitudinal Study of Ultrasound Biometric Measurements and Estimated Fetal Weight. \u003cem\u003ePLoS medicine \u003c/em\u003e2017, 14(1):e1002220.\u003c/li\u003e\n\u003cli\u003eOyelese Y, Vintzileos AM. The uses and limitations of the fetal biophysical profile. \u003cem\u003eClinics in perinatology \u003c/em\u003e2011, 38(1):47-64, v-vi.\u003c/li\u003e\n\u003cli\u003eBaschat AA, Galan HL, Lee W, DeVore GR, Mari G, Hobbins J, Vintzileos A, Platt LD, Manning FA. The role of the fetal biophysical profile in the management of fetal growth restriction. \u003cem\u003eAmerican journal of obstetrics and gynecology \u003c/em\u003e2022, 226(4):475-486.\u003c/li\u003e\n\u003cli\u003eBattarbee AN, Sandoval G, Grobman WA, Reddy UM, Tita ATN, Silver RM, El-Sayed YY, Wapner RJ, Rouse DJ, Saade GR\u003cem\u003e et al\u003c/em\u003e\u003cem\u003e.\u003c/em\u003e Maternal and Neonatal Outcomes Associated with Amniotomy among Nulliparous Women Undergoing Labor Induction at Term. \u003cem\u003eAmerican journal of perinatology \u003c/em\u003e2021, 38(S 01):e239-e248.\u003c/li\u003e\n\u003cli\u003eSmyth RM, Alldred SK, Markham C. Amniotomy for shortening spontaneous labour. \u003cem\u003eThe Cochrane database of systematic reviews \u003c/em\u003e2013(1):CD006167.\u003c/li\u003e\n\u003cli\u003eACOG Committee Opinion No. 766. Approaches to Limit Intervention During Labor and Birth. \u003cem\u003eObstetrics and gynecology \u003c/em\u003e2019, 133(2):e164-e173.\u003c/li\u003e\n\u003cli\u003eCallahan EC, Lee W, Aleshi P, George RB. Modern labor epidural analgesia: implications for labor outcomes and maternal-fetal health. \u003cem\u003eAmerican journal of obstetrics and gynecology \u003c/em\u003e2023, 228(5S):S1260-S1269.\u003c/li\u003e\n\u003cli\u003eAnim-Somuah M, Smyth RM, Cyna AM, Cuthbert A. Epidural versus non-epidural or no analgesia for pain management in labour. \u003cem\u003eThe Cochrane database of systematic reviews \u003c/em\u003e2018, 5(5):CD000331.\u003c/li\u003e\n\u003cli\u003eYi J, Chen L, Meng X, Chen Y. The infection, cervical and perineal lacerations in relation to postpartum hemorrhage following vaginal delivery induced by Cook balloon catheter. \u003cem\u003eArchives of gynecology and obstetrics \u003c/em\u003e2023.\u003c/li\u003e\n\u003cli\u003eMei-Dan E, Walfisch A, Suarez-Easton S, Hallak MJTJoM-F, Medicine N. Comparison of two mechanical devices for cervical ripening: a prospective quasi-randomized trial. 2012, 25(6):723-727.\u003c/li\u003e\n\u003cli\u003eTilden EL, Caughey AB, Lee CS, Emeis C. The Effect of Childbirth Self-Efficacy on Perinatal Outcomes. \u003cem\u003eJournal of obstetric, gynecologic, and neonatal nursing : JOGNN \u003c/em\u003e2016, 45(4):465-480.\u003c/li\u003e\n\u003cli\u003eFenwick J, Toohill J, Gamble J, Creedy DK, Buist A, Turkstra E, Sneddon A, Scuffham PA, Ryding EL. Effects of a midwife psycho-education intervention to reduce childbirth fear on women\u0026apos;s birth outcomes and postpartum psychological wellbeing. \u003cem\u003eBMC pregnancy and childbirth \u003c/em\u003e2015, 15:284.\u003c/li\u003e\n\u003cli\u003eOrganization WHO. WHO recommendations on intrapartum care for a positive childbirth experience: World Health Organization; 2018.\u003c/li\u003e\n\u003cli\u003eWanyenze EW, Byamugisha JK, Tumwesigye NM, Muwanguzi PA, Nalwadda GK. A qualitative exploratory interview study on birth companion support actions for women during childbirth. \u003cem\u003eBMC pregnancy and childbirth \u003c/em\u003e2022, 22(1):63.\u003c/li\u003e\n\u003cli\u003eJohansson M, Fenwick J, Premberg A. A meta-synthesis of fathers\u0026apos; experiences of their partner\u0026apos;s labour and the birth of their baby. \u003cem\u003eMidwifery \u003c/em\u003e2015, 31(1):9-18.\u003c/li\u003e\n\u003cli\u003eQin J, Liu X, Sheng X, Wang H, Gao S. Assisted reproductive technology and the risk of pregnancy-related complications and adverse pregnancy outcomes in singleton pregnancies: a meta-analysis of cohort studies. \u003cem\u003eFertility and sterility \u003c/em\u003e2016, 105(1):73-85.e71-76.\u003c/li\u003e\n\u003cli\u003eRozdarz KM, Flatley CJ, Kumar S. Intrapartum and neonatal outcomes in singleton pregnancies following conception by assisted reproduction techniques. \u003cem\u003eAust N Z J Obstet Gynaecol \u003c/em\u003e2017, 57(6):588-592.\u003c/li\u003e\n\u003cli\u003eSha T, Yin X, Cheng W, Massey IY. Pregnancy-related complications and perinatal outcomes resulting from transfer of cryopreserved versus fresh embryos in vitro fertilization: a meta-analysis. \u003cem\u003eFertility and sterility \u003c/em\u003e2018, 109(2):330-342.e339.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-pregnancy-and-childbirth","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"prch","sideBox":"Learn more about [BMC Pregnancy and Childbirth](http://bmcpregnancychildbirth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/prch/default.aspx","title":"BMC Pregnancy and Childbirth","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"machine learning, cephalic dystocia, risk assessment, prediction model","lastPublishedDoi":"10.21203/rs.3.rs-4776419/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4776419/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cb\u003eBackground\u003c/b\u003e\u003c/p\u003e \u003cp\u003eEarly detection of cephalic dystocia is challenging, and current clinical assessment tools are limited. Machine learning offers unique advantages, enabling the generation of predictive models using various types of clinical data. Our model aims to integrate objective ultrasound data with psychological and sociological characteristics and obstetric treatment data to predict the individual probability of cephalic dystocia in pregnant women.\u003c/p\u003e\u003cp\u003e\u003cb\u003eMethods\u003c/b\u003e\u003c/p\u003e \u003cp\u003eWe collected data from 302 pregnant women who underwent examinations and deliveries at Southern Medical University's Nanfang Hospital from January 2022 to December 2023. We utilized basic patient characteristics, foetal ultrasound parameters, maternal anthropometric data, maternal psychological measurements, and obstetric medical records to train and test the machine learning models. Our study analysed the effectiveness of three machine learning models: least absolute shrinkage and selection operator (LASSO) regression, decision tree, and random forest. The precision, accuracy, recall, and area under the receiver operating characteristic (ROC) Curve (AUC) were used to evaluate the performance of the models.\u003c/p\u003e\u003cp\u003e\u003cb\u003eResults\u003c/b\u003e\u003c/p\u003e \u003cp\u003eAmong the three machine learning models, the LASSO-based logistic regression model demonstrated the best predictive performance, with an AUC value of 0.833. We found that maternal ischial spine diameter, fetal biparietal diameter, fetal biophysical profile score, artificial rupture of membranes, labor analgesia, childbirth self-efficacy, and other variables were predictive factors for cephalic dystocia.\u003c/p\u003e\u003cp\u003e\u003cb\u003eConclusions\u003c/b\u003e\u003c/p\u003e \u003cp\u003eThis study constructed and validated a prediction model for cephalic dystocia via three machine learning methods, which can help clinicians improve the probability of identifying pregnant women at risk for cephalic dystocia.\u003c/p\u003e","manuscriptTitle":"Development and validation of a machine learning model for prediction of cephalic dystocia","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-08-19 23:52:46","doi":"10.21203/rs.3.rs-4776419/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-12-02T07:43:41+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-11-27T02:31:16+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-11-11T21:51:10+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"45742788351616361510772383931934124606","date":"2024-10-25T13:20:54+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"228276382073676875292032781289694742282","date":"2024-10-25T12:27:03+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-09-05T18:47:55+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"116576531552466592870396702153208683425","date":"2024-09-05T12:07:43+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-09-05T10:39:57+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-07-22T08:44:51+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-07-22T08:43:07+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Pregnancy and Childbirth","date":"2024-07-21T10:40:47+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-pregnancy-and-childbirth","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"prch","sideBox":"Learn more about [BMC Pregnancy and Childbirth](http://bmcpregnancychildbirth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/prch/default.aspx","title":"BMC Pregnancy and Childbirth","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9c80f9ca-c207-4ac3-9749-6f09290889b7","owner":[],"postedDate":"August 19th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-08-25T16:41:00+00:00","versionOfRecord":{"articleIdentity":"rs-4776419","link":"https://doi.org/10.1186/s12884-025-07972-8","journal":{"identity":"bmc-pregnancy-and-childbirth","isVorOnly":false,"title":"BMC Pregnancy and Childbirth"},"publishedOn":"2025-08-18 16:29:45","publishedOnDateReadable":"August 18th, 2025"},"versionCreatedAt":"2024-08-19 23:52:46","video":"","vorDoi":"10.1186/s12884-025-07972-8","vorDoiUrl":"https://doi.org/10.1186/s12884-025-07972-8","workflowStages":[]},"version":"v1","identity":"rs-4776419","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4776419","identity":"rs-4776419","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}