AbstractBackground Vaginal birth after cesarean section（VBAC） is recommended by international and domestic guidelines or expert consensuses.However ,no valid tools can exactly predict who can succeed in trying vaginal birth among eligible women with a history of cesarean section.Machine learning is gradually used to develop models in obstetrics and midwifery.This study aimed to develop an explainable machine learning model to predict the chance of successful VBAC. Methods The data were collected to establish 7 predicting models from two tertiary hospitals in Guangdong province of China.Training and internal validation data were collected from the First Dongguan Affiliated Hospital Of Guangdong Medical University from January 2012 to December 2022.External validation data were collected from Shenzhen Longhua District Central Hospital from Januray 2011 to December 2017. 7 predicting models based on machine learning were developed and evaluated by area under the operating characteristic curve (AUC).The optimal one was picked out from 7 models according to its AUC and other indices.The outcome of the predictive model was interpreted by Shapley Additive exPlanations(SHAP). Results A total of 2438 pregnant women with trial of labor after cesarean （TOLAC）were included in the final cohort. The CatBoost model was selected as the predictive model with the greatest AUC for 0.725 (95% CI: 0.653–0.792), the accuracy for 0.611 (95% CI: 0.557–0.672), sensitivity 0.69 (95% CI: 0.551–0.829), and specificity 0.69 (95% CI: 0.72–0.76). Cervical Bishop score and interval of pregnancy showed the greatest impact on successful vaginal birth, according to SHAP results. Conclusion Models based on machine learning algorithms can be used to predict whether a trail of vaginal birth succeeds. CatBoost model showed more significant performance compared with traditional logistic regression and other machine learning algorithms in this study. Cervical Bishop score and interval of pregnancy are important factors for successful VBAC. More researchs still need to be undertaken to promote the accuracy of ML algorithms and overcome their shortcomings.
Background: To address the worldwide dramatically increased caesarean section (CS) rate in the past decades, the World Health Organization (WHO) has recommended the CS rate should not be higher than 10-15%. Whether it is achievable remains unknown.Methods: We collected data of delivery from 2008 to 2017 in two typical regional hospitals in China: Longhua Hospital (national policies rigorously implemented) and Dongguan Hospital (national policies not rigorously implemented). We compared between two hospitals the 10 years trend in annual rate of CS, standardized by parity and CS history, against the time of issuing relevant national, local, and hospital policies.Findings: 42441 women in Longhua and 36935 women in Dongguan gave birth in the 10 years. China’s first national policy on CS reduction was issued in 2010 and the formal relaxation of one-child policy was in 2015-16. The standardized annual CS rate was around 40% in 2008-09, declined sharply since 2010 down to 10.9% in 2016 (p for trend <0.001) and leveled off then in Longhua. In contrast, the rate stayed around 30% at the beginning, started to decrease slightly since 2012 down to 26% in 2015 (p for trend <0.001), and then bounced back to 31% in 2017 in Dongguan. Proportion of women with history of CS increased significantly in two hospitals (both roughly from 6% before 2010 to 20% after 2015). Analyses stratified by modified Robson classification showed that CS rates reduced in all risk classes of delivery women in Longhua but only in the Robson class 2 group in Dongguan. Major complications did not differ by hospital.Interpretation: With vigorously implementing national policies at micro levels, the WHO recommended CS rate could be achieved without increase in major complications.Funding Information: The study was funded by Guangdong Provincial Science and Technology Plan (2017A020214007) and Dongguan Technology Development Society (201950715032188).Declaration of Interests: The author(s) declare that they have no competing interests.Ethics Approval Statement: The study was approved by the institutional review boards of the two participating hospitals. Informed consent was not obtained as this was a retrospective study and the data came from the hospital's medical records.
Background: To address the worldwide dramatically increased Cesarean section (CS) rate in the past decades, WHO has recommended the CS rate should not be higher than 10-15%. Whether it is achievable remains unknown. Methods: We collected the data of delivery from 2008 to 2017 in two typical regional hospitals in China: Longhua Hospital (national policies rigorously implemented) and Dongguan Hospital (national policies not rigorously implemented). We compared between the two hospitals the 10 years trend in annual rate of CS, standardized by age, education level, parity, and CS history, against the time of issuing relevant national, local, and hospital policies. Results: In 10 years, 42,441 women in Longhua and 36,935 women in Dongguan have given birth. China's first national policy on CS reduction was issued in 2010 and the formal relaxation of one-child policy was issued in 2015-2016. In Longhua, the standardized annual CS rate was around 35% in 2008-2009, which declined sharply since 2010 down to 13.1% in 2016 (p < 0.001) and then leveled off. In contrast, in Dongguan, the rate stayed around 25% at the beginning, increased to 36% in 2011, decreased sharply to 27% in 2012, and leveled off until 2015 (p < 0.001), and then bounced back to 35% in 2017. The proportion of women with the history of CS increased significantly in the two hospitals (both roughly from 6% before 2010 to 20% after 2015). Analyses stratified by modified Robson classification showed that CS rates reduced in all risk classes of delivery women in Longhua but only in the Robson class 2 group in Dongguan. Major complications did not differ by hospital. Conclusion: With vigorously implementing national policies at micro levels, the WHO-recommended CS rate could be achieved without increase in major complications.
Objectives To develop a nomogram to predict the likelihood of vaginal birth after caesarean section (VBAC) among women after a previous caesarean section (CS). Design A retrospective cohort study. Setting Two secondary hospitals in Guangdong Province, China. Participants Inclusion criteria were as follows: pregnant women with singleton fetus, age ≥18 years, had a history of previous CS and scheduled for trial of labour after caesarean delivery (TOLAC). Patients with any of the following were excluded from the study: preterm labour (gestational age <37 weeks), two or more CSs, contradictions for vaginal birth, history of other uterine incision such as myomectomy, and incomplete medical records. Primary outcome measure The primary outcome was VBAC, which was retrospectively abstracted from computerised medical records by clinical staff. Results Of the women who planned for TOLAC, 84.0% (1686/2006) had VBAC. Gestational age, history of vaginal delivery, estimated birth weight, body mass index, spontaneous onset of labour, cervix Bishop score and rupture of membranes were independently associated with VBAC. An area under the receiver operating characteristic curve (AUC) in the prediction model was 0.77 (95% CI 0.73 to 0.81) in the training cohort. The validation set showed good discrimination with an AUC of 0.70 (95% CI 0.60 to 0.79). Conclusions TOLAC may be a potential strategy for decreasing the CS rate in China. The validated nomogram to predict success of VBAC could be a potential tool for VBAC counselling.
Background: To address the worldwide dramatically increased caesarean section (CS) rate in the past decades, the World Health Organization (WHO) has recommended the CS rate should not be higher than 10-15%. Whether it is achievable remains unknown.Methods: We collected data of delivery from 2008 to 2017 in two typical regional hospitals in China: Longhua Hospital (national policies rigorously implemented) and Dongguan Hospital (national policies not rigorously implemented). We compared between two hospitals the 10 years trend in annual rate of CS, standardized by parity and CS history, against the time of issuing relevant national, local, and hospital policies.Fingdings:42441 women in Longhua and 36935 women in Dongguan gave birth in the 10 years. China’s first national policy on CS reduction was issued in 2010 and the formal relaxation of one-child policy was in 2015-16. The standardized annual CS rate was around 40% in 2008-09, declined sharply since 2010 down to 10.9% in 2016 (p for trend <0.001) and leveled off then in Longhua. In contrast, the rate stayed around 30% at the beginning, started to decrease slightly since 2012 down to 26% in 2015 (p for trend <0.001), and then bounced back to 31% in 2017 in Dongguan. Proportion of women with history of CS increased significantly in two hospitals (both roughly from 6% before 2010 to 20% after 2015). Analyses stratified by modified Robson classification showed that CS rates reduced in all risk classes of delivery women in Longhua but only in the Robson class 2 group in Dongguan. Major complications did not differ by hospital.Interpretation: With vigorously implementing national policies at micro levels, the WHO recommended CS rate could be achieved without increase in major complications.Funding: The study was funded by Guangdong Provincial Science and Technology Plan (2017A020214007) and Dongguan Technology Development Society (201950715032188).Declaration of Interest: The author(s) declare that they have no competing interests.Ethical Approval: The study was approved by the institutional review boards of the two participating hospitals. Informed consent was not obtained as this was a retrospective study and the data came from the hospital's medical records.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)