Prediction of odds for emergency cesarean section: A secondary analysis of the CHILD term birth cohort study.

INTRODUCTION: Previously developed cesarean section (CS) and emergency CS prediction tools use antenatal and intrapartum risk factors. We aimed to develop a predictive model for the risk of emergency CS before the onset of labour utilizing antenatal obstetric and non-obstetric factors.
METHODS: We completed a secondary analysis of data collected from the CHILD Cohort Study. The analysis was limited to term (≥37 weeks), singleton pregnant women with cephalic presentation. The sample was divided into a training and validation dataset. The emergency CS prediction model was developed in the training dataset and the performance accuracy was assessed by the area under the receiver operating characteristic curve(AUC) of the receiver operating characteristic analysis (ROC). Our final model was subsequently evaluated in the validation dataset.
RESULTS: The participant sample consisted of 2,836 pregnant women. Mean age of participants was 32 years, mean BMI of 25.4 kg/m2 and 39% were nulliparous. 14% had emergency CS delivery. Each year of increasing maternal age increased the odds of emergency CS by 6% (adjusted Odds Ratio (aOR 1.06,1.02-1.08). Likewise, there was a 4% increase odds of emergency CS for each unit increase in BMI (aOR 1.04,1.02-1.06). In contrast, increase in maternal height has a negative association with emergency CS. The final emergency CS delivery predictive model included six variables (hypertensive disorders of pregnancy, antenatal depression, previous vaginal delivery, age, height, BMI). The AUC for our final prediction model was 0.74 (0.72-0.77) in the training set with a similar AUC in the validation dataset (0.77; 0.71-0.82).
CONCLUSION: The developed and validated emergency CS delivery prediction model can be used in counselling prospective parents around their CS risk and healthcare resource planning. Further validation of the tool is suggested.

Introduction

The World Health Organization (WHO) has raised concerns regarding the dramatic increase in cesarean section (CS) rates. CS is effective in managing dystocia and other significant complications of pregnancy. Indications for scheduled CS can be divided into absolute and relative indications [1]. Absolute indications for scheduled CS include cephalopelvic disproportion, placenta previa, abnormal lie and presentation. Prior CS delivery is a relative indication for scheduled CS [1-5].

Emergency CS is indicated when acute obstetrical complications that threaten the life of the mother and/ or the fetus including fetal distress and antepartum hemorrhage develop. Intrapartum factors such as labour dystocia, fetal distress, and umbilical cord prolapse are absolute indications for emergency CS [1, 6, 7]. Emergency CS, is associated with increased maternal morbidity and mortality, compared to a scheduled CS. Morbidity associated with emergency CS include severe hemorrhage, complications from rapid administration of general anesthesia and accidental injury to the mother and infant [8-11]. A meta-analysis reported that the rates of maternal and fetal complications and mortality were higher in emergency CS when compared to scheduled CS [12]. In addition to the additional morbidity and mortality, resource planning for an emergency CS is more difficult compared to scheduled CS resulting in higher infection rates [9].

The CS risk prediction model developed by Janssen et al and Souza et al utilized both antenatal and intrapartum factors for low risk nulliparous pregnant women [13, 14]. The FLAMM scoring system, developed to predict a VBAC (vaginal birth after prior cesarean section), included intrapartum factors including cervical dilation and effacement. The Grobman calculator [15], which included only antenatal factors, has limited generalizability as the tool is meant to predict the probability of a vaginal birth after cesarean section (VBAC) for term pregnant women with one prior CS. Tools to predict emergency CS delivery have incorporated antepartum and intrapartum factors [16, 17]. The emergency CS risk prediction model and classification tree (CTREE), with discriminatory accuracy ranges from 0.74 to 0.81, included intrapartum factors such as scalp pH, and labour induction among women with history of previous CS [18]. A risk scoring system for emergency CS was developed utilizing both antenatal and intrapartum factors such as quantity and characteristic of the amniotic fluid in Chinese population [3]. We could not identify a tool or scoring system for emergency CS risk prediction utilizing prenatal factors only. In this study, we used data from the CHILD Cohort Study to identify the main antenatal obstetric and non-obstetric risk factors for emergency CS and to subsequently develop an emergency CS prediction tool.

Materials and methods

Study design and participants

This was a secondary analysis of the CHILD Cohort Study, a large general-population recruited prospective observational pre-birth cohort study of 3,455 pregnant women enrolled in Edmonton, Winnipeg, rural Manitoba, Vancouver and Toronto between 2009 and 2012. This secondary analysis focused primarily on the emergency CS prediction. Details on the data collection methods and the characteristics of the cohort have been described previously [19] (www.childstudy.ca). Mothers were approached for enrollment in the study during the second or third trimester of their pregnancy. Infants, and their parents, were recruited if born at 34 weeks’ gestation or later and with birth weight of 2,500 g or more.

Mothers completed questionnaires on general health such as diabetes, hypertension and psychosocial factors at the time of recruitment and at 36 weeks of gestation. Information regarding maternal age, weight (kg), height (cm), parity, socioeconomic status, maternal education, ethnicity, maternal smoking status, medical comorbidities and risk factors including hypertensive disorder [20] and diabetes mellitus complicating pregnancy [21] were collected through standardized questionnaire. Maternal antenatal depression was assessed using the Center for Epidemiologic Studies Depression Scale (CES-D) [22]. Participants were classified as depressed if their CESD-score was ≥10 points. The socioeconomic status (SES) was divided into two groups with a cut-off income of ≥ $60,000 which indicates higher socioeconomic status. Maternal early pregnancy BMI was calculated using self-reported height and weight and classified by World Health Organization (WHO) criteria. Delivery information, including delivery mode, gestational age at birth and neonatal sex, were obtained from birth chart reviews. Pregnant women provided written informed consent to participate in the CHILD study. Ethics approval was obtained from the research ethics board of each CHILD study center (University of Alberta research ethics board, McMaster University research ethics board, Hospital for Sick Children research ethics board, University of Manitoba research ethics board, and University of British Columbia research ethics board) in addition to the McMaster University research ethics board. Patients provided informed written consent to have data from their medical records used for the CHILD study. A separate ethics approval was obtained from the University of Alberta Research Ethics Board for this secondary data analysis (Pro00092920). The data used for this study was de-identified (no participant were identifiers included in the dataset) prior to being released for analysis.

The present study restricted the analysis to nulliparous and multiparous women carrying a singleton, cephalic presentation fetus at 37 completed weeks of gestation with available birth chart records (n = 3,408). The remaining exclusion criteria were: women with a higher risk of scheduled CS such as placenta previa, a prior CS delivery, multiple gestation, cephalopelvic disproportion, breech presentation, pre-term, home birth and those who had their labour induced.

Statistical analysis

Data analysis steps to develop an emergency CS score are described in . A total of 2,836 pregnant women met the inclusion criteria. First, the data were randomly divided into two groups: a training dataset (80% of the sample, n = 2,269) and a validation dataset (20% of the sample, n = 567). The demographic, antenatal and obstetric characteristics of training and validation data set was shown in S1 Table in . Categorical variables were analyzed using the Chi-squared test or the Fisher’s exact test and t-test was employed for the continuous variables. All parametric data were expressed as mean ± standard deviation (SD), and non-parametric data as median ± interquartile range (IQR). The primary outcome was emergency CS for any indication. Mean centering was employed to center the maternal age and height variables.

In the training dataset, the univariate and multiple logistic regression models were used to determine the factors associated with emergency CS. The predictors considered for the model included maternal age, ethnicity, height, weight, BMI, gestational age at delivery and parity. Variable selection for the CS risk prediction was based on combination of literature review (S2 Table and S1 Appendix in , clinical experience and found to be significant in univariate analysis. The variables with a p-value of <0.20 in univariate analyses were included in the multiple logistic regression model. A prediction model (vaginal vs. emergency CS) was then developed with the training data set taking hospital or province difference of cesarean section rate into account. The C-statistic, area under Receiver operating characteristic curve (AUC), was used to assess the performance of the prediction model based on the model’s sensitivity and specificity. The final model was adjusted for maternal height, BMI, CESD-score, hypertensive disorders of pregnancy, history of previous vaginal delivery and hospital CS rate.

The predictive ability of the model was then evaluated in the validation data set. The p-values for all hypothesis tests were 2-sided and statistical significance was set at p <0.05 for all analyses. Goodness-of-fit for the logistic regression models was assessed by using the Hosmer and Lemeshow test. The scoring system was developed based on the weighted estimate of the multiple logistic regression model. The flow diagram of the statistical analysis was described in S1 Fig in . Data analysis was carried out using STATA version 14.

Results

The demographic and clinical characteristic of women in the study cohort are presented in . Of the 2,836 low-risk pregnant women included in the final analysis, 22% had a CS delivery with 14% (365/2680) emergency CS delivery. The majority of women enrolled were Caucasians (73%). The mean age of women at enrollment was 32 years with a mean BMI of 25 kg/m2. Among infants delivered by emergency CS, 59% (214/365) were male. Among the women delivered by emergency CS, 6% had gestational diabetes, 7% had hypertensive disorders of pregnancy and 20% of the babies were delivered before reaching full-term (≥ 39 weeks). Women delivered by emergency CS had greater depression symptoms (CESD-scores ≥ 10 points) than the vaginally delivered group (35% vs. 27%, p = 0.0001). The women with emergency CS were older and had higher BMI when compared to vaginally delivered women (33.70 vs. 31.99, p = 0.01; 32.63 vs. 31.99, p = 0.023). S1 Table in shows the clinical characteristics of considered predictors in the training and validation data set.

In multiple logistic regression, women with antenatal depression score ≥ 10 points had a 45% increased risk of being delivered by emergency CS (aOR 1.45, 1.07–1.96). Each year of increasing maternal age increased the odds of CS by 6% (aOR 1.06, 1.02–1.08) and each unit increase in BMI increased the odds of CS by 4% (aOR 1.04, 1.02–1.06). Pregnant women who had an emergency CS were more likely to have hypertensive disorders of pregnancy (aOR 1.75, 1.01–3.07). In contrast, taller pregnant women (aOR 0.94, 0.92–0.96) and women who had a previous vaginal delivery had lower odds of having an emergency CS (aOR 0.21, 0.15–0.30). Women who had history of previous vaginal delivery (aOR 0.46, 0.36–0.59) was significantly associated with decreased risk of CS (. In our stratified analysis by parity, hypertensive disorders of pregnancy was a significant predictor for CS in nulliparous but not multiparous mwomen (aOR 1.93, 1.02–3.67 vs. aOR 1.09, 0.55–2.21) (S3 Table in . We also conducted sensitivity analysis with the exclusion of CESD-score variable from our prediction model (S4 Table in and the findings are comparable to the full model.

Our emergency CS model identified six predictors when controlling for hospital delivered: maternal age, height, BMI, hypertensive disorders of pregnancy, antenatal depression score (CES-D), previous history of vaginal delivery (. The AUC values for the development prediction models was 0.74 (0.72–0.77) while the AUC for the validation dataset was 0.77 (0.71–0.82) (. The calibration curve of the prediction model was presented in . We subsequently developed a modified scoring system based on the logistic regression model coefficients that ranged from 0 to 14 (. The scores were further categorized into grade 0 (0–5 points), grade 1 (6–7 points), grade 2 (8–9 points), and grade 3 (≥ 10 points). With the increase in grade, there was an increase in odds of emergency CS risk (. For example, individuals with grade 2 had a 6.11 increased odds of having an emergency CS (95%CI; 3.06–12.19) while individuals with grade 3 had a 13.96 increased odds of an emergency CS (95%CI; 7.32–26.61) compared to individuals with grade 0 (baseline) risk. The developed modified scoring system provided a sensitivity of 11%, specificity of 91% and an AUC of 0.70 (0.68–0.73) (. Among women with a grade 1 risk of an emergency CS, the number needed to treat (NNT) is seven (i.e. schedule seven CS to prevent one emergency CS), while the NNT was three for emergency CS grade 2 while NNT = 4 and women with a grade 3 emergency CS risk. We also developed an emergency CS risk prediction calculator in the Redcap which can be utilized by the healthcare professionals to assess the risk and provide counselling to the high risk pregnant women.

Discussion

We developed a score that identifies low-risk pregnant women at risk for an emergency cesarean using data from a large population based cohort from different sites in Canada. The score includes antenatal obstetric and non-obstetric factors, as well as birth order of the infant, and controls for the hospital CS rate. The yielded AUC are comparable to prediction models that included intrapartum factors [14, 23], birth weight of the infants [23] and premature rupture of membrane [23]. Most of the parameters in our predictive model are routinely collected as part of routine prenatal care except the CES-D (maternal depression) score. Furthermore, our model has good generalizability as the score was developed from deliveries from 13 different hospitals distributed across Canada. The emergency CS scores could be utilized in the overall context of clinical information to help patient with counseling, expectation and decision-making.

Several studies, including our own, have shown that advanced maternal age was associated with higher odds of having a CS delivery [14, 23–25]. Similarly, our finding of an inverse association between maternal height and CS delivery is consistent with prior studies [12, 21, 24, 26]. Furthermore, a higher maternal BMI has been associated with adverse obstetric outcome and increased the risk of CS delivery [27, 28]. A previous history of vaginal delivery decreased the risk of emergency CS were consistent with the findings from VBAC prediction models [24, 29]. In contrast to prior studies, we did not find that sociodemographic factors such as ethnicity, education and social class and employment and income status were associated with emergency CS [17, 30]. The CS prediction models from Souza et al, a multicenter study, included both maternal and fetal antenatal and intrapartum factors such as cervical position, fetal station, fetal distress and fetal head molding [13]. The AUC of our prediction model containing only antenatal factors (074, training, 0.77 validation) was comparable to the model developed by Souza et al (AUC 0.78). The emergency CS risk scoring system by Guan et al (2020) [3], with an AUC of 0.79, included intrapartum factors such as quantity and color of the amniotic fluid.

Similar to the findings from the previous studies [31-34], we found that multiparous with a history of previous vaginal delivery and with cephalic presentation had lower risk of CS delivery. Hypertensive disorders of pregnancy increased the risk of CS delivery [35]. Nulliparous women have significant higher risk of developing hypertensive related disorders than multiparous women [36, 37]. Similarly, we found pregnancy-induced hypertension increased the risk of emergency CS among nulliparous but not multiparous pregnant women. Nonetheless, unforeseen circumstances such as prolonged labour and fetal distress can occur in multiparous women with prior vaginal delivery. Emergency CS is indicated when acute complications like fetal distress and antepartum hemorrhage develop and threaten the life of the mother and/ or the fetus. Careful assessment and monitoring during antenatal and intranatal period should be provided to both nulliparous and multiparous to improve maternal and neonatal outcome. Our prediction tool specifically excluded intra-partum factors associated with emergency CS such as labour dystocia and fetal distress. The contribution of labour dystocia and fetal distress on emergency CS rates are increasing area of focus for intervention. The American College of Obstetricians and Gynecologists (ACOG) guidelines for reducing the primary CS due to labour dystocia [1] were developed in 2014. As such, these guidelines were not implemented in the Society of Obstetricians and Gynaecologists of Canada (SOGC) guidelines [38] during the study recruitment period (2009–2012) [1]. Further work will examine the role of our tool in the context SOGC guidelines for management of labour dystocia and fetal distress.

The components of our eCS score including maternal age, height, BMI, childbirth history and hypertensive disorders of pregnancy are often collected as part of routine prenatal care. Additionally, unlike prior emergency CS tools, our score does not include any intrapartum data allowing for application at any point during pregnancy. The antenatal obstetric and non-obstetric factors identified from our prediction tool can be utilized in screening and identification of individuals at high risk for an emergency CS. Increase surveillance and antennal interventions could be provided for the modifiable antenatal risk factors such as low dose aspirin for hypertension, counseling for depression and weight management for overweight pregnant women.

Unique to our study, we observed that women with higher antenatal depression score had higher risk of emergency CS delivery. One study reported that mental health status, in particular stress, sleep disturbances and worry were associated with higher risk of emergency CS [39]. Fear and anxiety of childbirth [40, 41] and depressed mood [42] are common causes for preference for CS. Our study finding suggests clinicians should assess for the presence of antenatal depression in routine antenatal screening for emergency CS risk. In addition, comprehensive mental health programs and the effective interventions of health promotion could reduce the fear and promote confidence with childbirth by vaginal delivery. We were not able to develop a scoring system for scheduled CS with a significant predictive capacity. The Avon Longitudinal Study reported that the largest impact on scheduled CS was breech presentation and previous CS [17]. Our exclusion of women with breech presentation, prior CS delivery, placenta previa and cephalopelvic disproportion, and abnormal lie and presentation from the analysis, known risk factors for a scheduled CS [4], may have resulted in the inability of our scoring system to predict the risk of scheduled CS. Additionally, the data on prior CS history is incomplete in our study population. Hence, we cannot be certain whether the observed increased in the risk of scheduled CS in the subsequent born children could be a confounding effect of prior CS history.

Strengths of our study include a nationwide, prospective design, conducted in a large birth cohort study from four sites in Canada. With the multinomial logistic regression model, the risk of emergency CS were estimated and the parameter estimates are more efficient with less error. Our study had access to the wide range of sociodemographic and pregnancy related variables beyond what would normally be available in a clinical chart review. In addition, many of the antenatal factors utilized for the prediction model were verified with birth chart review by research assistant. Finally, the large sample size provided us with sufficient power to predict emergency CS risk and develop the scoring system with internal validation.

Our research is not without limitation. The observational study design with self-reported items may introduce systematic error in the variance of the predictor variables. Our study did not have access to complete information on maternal weight change during pregnancy, presence of oligohydramnios and estimated fetal weight. We only included term infants in this analysis as the risk factors for CS are different in pre-term infants. Our prediction model included both nulliparous and multiparous pregnant women which may have impact on model. Nonetheless, we adjusted for birth order in our prediction model as well as undertaking sensitivity analysis in nulliparous and multiparous subgroups. We did not find that socioeconomic status impacted our findings. This may be the result of a higher socioeconomic status study sample or a reflection of our publicly-funded healthcare system. Our observed lower CS rate may be due to the lower proportion of overweight women in the study. Lack of information on estimated fetal weight during the third trimester may limit the prediction ability of our model.

While we performed internal validation by splitting the data set, we lacked data for conducting external validity for our CS prediction model. Future research could include external validation of the score in other large, prospectively cohort study. The lack of complete information on prior CS will be worth exploring as an explanation of the variation in scheduled CS and the role of women’s preferences. Subsequent work may assess the impact of our prediction model in decision-making about timing and mode of delivery and thereby influence acute and long-term outcomes for women and their offspring. The CESD-score used in the prediction model are not routinely collected during antenatal care. Further studies may consider validating our emergency CS prediction tool with routinely collected antenatal depression questions.

The high specificity and low sensitivity suggest that the tool is good at determining who will not need an eCS (a rule-out test). As such, we would recommend that women who screen positive have closer pre-natal follow-up. The tool identifies areas for potential intervention to lower an individual’s risk for an eCS. Our study indicated that women with a higher BMI were more likely to have an emergency CS delivery and. Future research will examine weight control efforts before and during pregnancy may help to reduce the emergency CS rate [27].

Conclusions

We successfully developed a model to predict the likelihood of emergency CS using prenatal obstetric and non-obstetric factors. The proposed prediction model has similar performance characteristics compared to other emergency CS prediction models without the need for intra-partum prediction factors. The tool could be used in conjunction with the Grobman VBAC tool [15] to assist in delivery mode decision-making and healthcare resource planning and allocation. Early identification of the women at an increased risk of emergency CS is important for patient management including referral for mental health counseling and weight management program to prevent emergency CS. Further prospective validation studies in the general population should be undertaken to confirm efficacy of the developed prediction model and the scoring system before application in the general population.

(DOCX)

Click here for additional data file.

1 Jul 2021

PONE-D-21-10805

Prediction of odds for emergency cesarean section: a secondary analysis of the CHILD term birth cohort study

PLOS ONE

Dear Dr. Mandhane,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Aug 15 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see:  http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at  https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Eduardo Ortiz-Panozo, MD; MSc

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for including your ethics statement: "Ethics approval was obtained from local authorized review board of each CHILD study center and McMaster University. A separate ethics approval was obtained for this secondary data analysis (Pro00092920)."

a) Please amend your current ethics statement to include the full name of the ethics committee/institutional review board(s) that approved your specific study.

c) Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified (1) whether consent was informed and (2) what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If your study included minors, state whether you obtained consent from parents or guardians. If the need for consent was waived by the ethics committee, please include this information.

If you are reporting a retrospective study of medical records or archived samples, please ensure that you have discussed whether all data were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have data from their medical records used in research, please include this information.

Once you have amended this/these statement(s) in the Methods section of the manuscript, please add the same text to the “Ethics Statement” field of the submission form (via “Edit Submission”).

For additional information about PLOS ONE ethical requirements for human subjects research, please refer to http://journals.plos.org/plosone/s/submission-guidelines#loc-human-subjects-research.

3. In ethics statement in the manuscript and in the online submission form, please provide additional information about the patient records/samples used in your study. Specifically, please ensure that you have discussed whether all data/samples were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have data/samples from their medical records used in research, please include this information.

4. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

5. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

Additional Editor Comments:

Since the study aims at developing a prediction model, I suggest to include a table showing the final model's regression coefficients plus the intercept. Not as a supplementary table (i.e, Supplementary table 3) but as part of the main analysis.

Please clarify how many and what variables were tested as potential predictors. What were the p-values for excluded variables?

Authors mention Table 3a and a scoring system in the text, but I could not find those results in tables.

I would suggest to include more statistical measures of the predictive capacity of the model. Is it possible to calculate diagnostic accuracy, sensitivity, specificity, PPV, NPV and/or diagnostic likelihood ratios?

It would be important that authors discuss about the differences between the models for nulliparous and multiparous and their implications. From their estimated coefficients, it seems like they only differ in the role of hypertensive disorders of pregnancy and (obviously) previous vaginal delivery.

The model is of limited utility for settings where there is no information on CES-D. Since most likely this would be the case for most situations, I suggest to run separate models without this variable.

How these results compare to other prediction model? I would suggest to comment about the WHO predictions models, such as C-Model by Souza et al.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The manuscript is relevant and reaches conclusions according to the objectives, but there is a lack of attemption to the analgesic modalities during labour and about the epidural rates. I think it is a very important issue when you are affording CS rates. Statistical analysis has been correctly performed and statistical material has been provided properly. The manuscript is written in correct english, as far as I know (I'm not native english speaker).

The authors should include details about analgesia during labour, epidural rate, and motor block, or at least to detail the epidural analgesic protocol that has been used in this cohort of women.

Reviewer #2: This is an interesting study.

However,

1. Some grammatical etc. errors should be corrected. Examples:

Line 110/Please change “earlypregnancy” to “early pregnancy”

Line 119/Please change “cephalopevlic” to cephalopelvic” (using the advantage of automatic corrections).

2. In the “conclusion”, a real conclusion is expected (and not a detailed iteration of the title).

3. To take advantage of the practical benefits of similar studies, it should be clarified:

a) the indications of emergency CS (ECS)

b) the necessary steps to avoid situations leading to ECS

4. A working calculator could convince that the whole effort is a useful tool to justify or avoid ECS.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

30 Sep 2021

We have provided a file in the response documenting our response to the editor and reviewer comments.

Submitted filename: 2021-08-15 PLOS One R1 Response To Reviewers.docx

Click here for additional data file.

8 Feb 2022

Prediction of odds for emergency cesarean section: a secondary analysis of the CHILD term birth cohort study

PLOS ONE

Dear Dr. Mandhane,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Mar 25 2022 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

21 Mar 2022

Review Comments to the Author

Reviewer #3: Overall, this study was interesting as they had a large number of samples and a combination of obstetrics and non-obstetrics factors counted for emergency cs prediction tools. However, there are a few things that could be added to improve the quality of the study.

In general, it was suggested to use appropriate written English language. Many paragraphs were considered not effective as repetitive words and redundant sentences were found.

- Abstract:

Keywords: antenatal depression?

Higher AUC in validation set?

Conclusion: Maybe you could paraphrase?

Response: Thank you for the suggestion. We have updated the abstract accordingly. Antenatal depression is synonymous with perinatal depression.

- Introduction

No reference 1? The reference started with number 2

Response: Updated the reference.

Line 64-66, 66-68: please paraphrase

Response: Rephrased the sentences (line 60-67)

Indications for scheduled CS can be divided into absolute and relative indications (1). Absolute indications for scheduled CS include cephalopelvic disproportion, placenta previa, abnormal lie and presentation. Prior CS delivery is a relative indication for scheduled CS ((1–5) .

Emergency CS is indicated when acute obstetrical complications that threaten the life of the mother and/ or the fetus including fetal distress and antepartum hemorrhage develop. Intrapartum factors such as labour dystocia, fetal distress, and umbilical cord prolapse are absolute indications for emergency CS (1,6,7)

- Methods & Results

When dividing the data? After recruitment or after applying exclusion criteria?

Response: The data were divided after applying the exclusion criteria.

Please specify with number too. The division of training and validation set (80% and 20%) was based on what?

Response: The division of training and validation data set (80% vs. 20%) was based on random splitting. Updated accordingly (line 131-134).

“Data analysis steps to develop an emergency CS score are described in Figure 1. A total of 2,836 pregnant women met the inclusion criteria. First, the data were randomly divided into two groups: a training dataset (80% of the sample, n=2,269) and a validation dataset (20% of the sample, n=567).”

The use of CESD in methods, please put a reference.

Response: Reference added (line 108).

Diagnosis criteria for hypertensive disorder and diabetes mellitus complicating pregnancy? please put reference too

Response: References added (line 106)

The Adjusted OR, please specify, adjusted with what?

Response: updated accordingly in the methods section (line 152-153)

“The final model was adjusted for maternal height, BMI, CESD-score, hypertensive disorders of pregnancy, history of previous vaginal delivery and hospital CS rate.”

- Analysis:

This study focuses in comparing vaginal birth and emergency cs. Since you put scheduled cs in the table 1, why don’t you also compare the scheduled cs and the emergency cs? The findings might be interesting and can be added in the discussion too.

Response: Thank you for the suggestion. The main purpose of this study is to develop an emergency CS risk prediction tool. The study excluded major indications for scheduled CS (i.e. cephalopelvic disproportion, placenta previa, previous CS delivery) in the analysis. Additionally, several tools for predicting CS already exist.

- Discussion:

Please add discussion to compare other studies for emergency cs scoring systems. Not only study for cs prediction in general. Such as a current previous study which develop a scoring system for emergency cs by maternal-fetal perinatal characteristics. Although that one was different from yours, maybe you can add this in the discussion.

Response: We updated with the emergency CS scoring system (line 253-254).

“The emergency CS risk scoring system by Guan et al (2020) (3), with an AUC of 0.79, included intrapartum factors such as quantity and color of the amniotic fluid.”

Typo: Line 144, 155, 249

Response: Typo errors corrected (lines 145, 158, 252)

- Reference:

A number of references were considered too old, there are many newer references that could be used related to the issue

Response: Updated the references

- Supplementary materials:

Table 1: acupressure? Reference?

Response: reference updated

Table 2 in supplementary files similar to Table 1 in the manuscript? No need to mention it in supplementary files.

Response: Table 1 in the manuscript shows the demographic, antenatal and obstetric characteristics associated with mode of delivery. Table 2 in supplementary files presents the demographic, antenatal and obstetric characteristics of Training and Validation data (line 134-135)

“The demographic, antenatal and obstetric characteristics of training and validation data set was shown in S2 Table. “

Reviewer #4: The authors have done a good job attending the reviews, especially the one related to the comparison with other tools predicting the risk of C-Section. However, I think it is still possible in the discussion to stress what are the advantages of this tool. The authors must make a strong case for why this new tool is necessary, and the advantages of using this one instead of other existing tools up front.

Response: We have updated accordingly. (line 274-277)

“The components of our eCS score including maternal age, height, BMI, childbirth history and hypertensive disorders of pregnancy are often collected as part of routine prenatal care. Additionally, unlike prior emergency CS tools, our score does not include any intrapartum data allowing for application at any point during pregnancy.”

Submitted filename: Rebuttal.docx

Click here for additional data file.

31 Mar 2022

20 Apr 2022

Thank you for the insightful comments and the opportunity to further improve our manuscript. We have provided responses to each of the editor’s comments and have amended the manuscript accordingly.

Query 1: I would request authors to comment on the implications of their tool's low sensitivity in the discussion. Does that mean the tool works better as a "rule-in" rather than a "rule-out" test? And if so, do the relatively high AUC and accuracy result not from identifying (predicting) those who need an emergency CS but those who would not need one?

Response: We agree that additional clarification of the tool will help the readers understand the tool’s applicability in practice. The high specificity and low sensitivity suggest that the tool is better at determining who will not need an eCS (a rule-out test). The editor is correct that the high AUC and accuracy are from identifying those who do not need an eCS (the majority of women). We would recommend that women who screen positive have closer pre-natal follow-up rather than scheduling a CS. The tool also identifies areas for potential intervention to lower an individuals risk for an eCS. We have added the following to the discussion (Lines 305 – 311)

“The high specificity and low sensitivity suggest that the tool is good at determining who will not need an eCS (a rule-out test). As such, we would recommend that women who screen positive have closer pre-natal follow-up. The tool identifies areas for potential intervention to lower an individual’s risk for an eCS.”

Query 2: Please also confirm that PPV in the validation sample is 63% (Table 2).

Response: We have confirmed that the PPV of the validation sample is 63%. We have attached the classification table for the validation dataset in this response (appendix 1 of the response).

Submitted filename: 2022-04-20 PLOS one R3 response.docx

Click here for additional data file.

26 Apr 2022

Prediction of odds for emergency cesarean section: a secondary analysis of the CHILD term birth cohort study

PONE-D-21-10805R3

Dear Dr. Mandhane,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Eduardo Ortiz-Panozo, MD; MSc

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

4 May 2022

PONE-D-21-10805R3

Prediction of odds for emergency cesarean section: a secondary analysis of the CHILD term birth cohort study

Dear Dr. Mandhane:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Eduardo Ortiz-Panozo

Academic Editor

PLOS ONE

Table 1

Demographic, antenatal and obstetric characteristics associated with mode of delivery.

Characteristics	Vaginal * (n = 2,315)	Emergency CS (n = 365)		Scheduled CS (n = 156)
Maternal Age (years) (mean ± SD)	31.99 ± 4.62	32.63 ± 4.86		33.70 ± 4.30
Maternal Height (cm) (mean ± SD)	165.53 ± 6.81	162.79 ± 6.98		164.60 ± 7.50
Maternal Weight (kg) (mean ± SD)	68.75 ± 16.42	70.90 ± 18.37		73.06 ± 20.36
BMI in kg/m2 (mean ± SD)	25.07 ± 5.68	26.71 ± 6.46		26.89 ± 6.89
Hospitals CS rate CHILD cohort (mean ± SD)	5.93 ± 3.20	6.57 ± 3.16		7.4 ± 3.30
Increased CESD-score (Ref: <10)	614 (27%)	128 (35%)		49 (31%)
Gestational Age (weeks)
37	133 (6%)	29 (8%)		9 (6%)
38	253 (11%)	42 (12%)		35 (23%)
39	551 (24%)	65 (18%)		75 (49%)
40	754 (33%)	92 (26%)		29 (18%)
41	514 (22%)	104 (29%)		5 (2.5%)
≥42	97 (4%)	27 (7%)		1 (0.5%)
Gravida
G1	862 (37%)	199 (55%)		38 (24%)
G2	748 (32%)	90 (25%)		56 (36%)
G3	386 (17%)	38 (10%)		34 (22%)
G4	176 (8%)	21 (6%)		14 (9%)
≥G5	142 (6%)	16 (4%)		14 (9%)
Maternal Ethnicity
Caucasian	1334 (74%)	157 (67%)		111 (74%)
Others	462 (26%)	77 (33%)		40 (26%)
Marital status
Married or Common Law	1982 (86%)	302 (83%)		136 (87%)
Single (Never been married)	113 (5%)	7 (5%)		17 (5%)
Divorced/Widowed/ Separated	220 (9%)	56 (12%)		3 (8%)
Socioeconomic status
<$60,000	418 (21%)	62 (19%)		18 (14%)
≥ $60,000	1592 (79%)	258 (81%)		115 (86%)
Maternal Education
No education beyond high school	209 (9%)	30 (9%)		6 (4%)
Some post secondary/ college	448 (20%)	87 (25%)		41 (28%)
University degree	1559 (71%)	235 (67%)		98 (68%)
Maternal smoking history
Yes	133 (7%)	19 (8%)		14 (9%)
Hypertensive Disorders of Pregnancy
Yes	69 (3%)	26 (7%)		6 (4%)
Gestation Diabetes
Yes	98 (4%)	23 (6%)		9 (6%)
Previous vaginal delivery
First Born	1170 (51%)	289 (79%)		47 (30%)
Subsequent Born	1141 (49%)	75 (21%)		109 (70%)
Child Sex
Male	1204 (52%)	214 (59%)		79 (51%)
Female	1111 (48%)	151 (41%)		77 (49%)
Analgesia
Epidural	1414 (61%)	275 (75%)		10 (6%)
Spinal	27 (1%)	90 (25%)		151 (97%)
General Anesthesia	5 (0.2%)	23 (6.3%)		5 (3%)

* = Vaginal delivery was used as a reference and compared with emergency CS and scheduled CS.

p-values <0.05 in bold; SD = standard deviation; BMI = body mass index, PE = preeclampsia.

Table 2

Multiple logistic regression results include demographic, antenatal physical and obstetric characteristics in overall cohort independent of the parity: (Training and validation dataset).

	Emergency CS (Training, n = 2150)			Emergency CS (Validation, n = 530)
	Coefficient	Odds Ratio	95% CI	Coefficient	Odds Ratio	95% CI
Centered Age (years)	0.05	1.06	1.02–1.08	0.13	1.14	1.07–1.22
Centered Height (cm)	-0.06	0.94	0.92–0.96	-0.06	0.94	0.90–0.98
BMI in kg/m2	0.04	1.04	1.02–1.06	0.07	1.07	1.03–1.12
CESD-score (ref: <10)	0.37	1.45	1.07–1.96	0.51	1.66	1.01–3.15
Hospital CS rate (CHILD)	0.04	1.04	0.98–1.09	0.16	1.17	0.98–1.28
Hypertensive Disorders of Pregnancy	0.56	1.75	1.99–3.08	0.28	1.32	0.36–4.80
Previous vaginal delivery	-1.57	0.21	0.15–0.29	-1.63	0.20	0.10–0.38
Constant	-2.85			-4.33
AUC		0.74	0.72–0.77		0.77	0.71–0.82
Sensitivity		12%			13%
Specificity		99%			98%
Positive Predictive Value		28%			63%
Negative Predictive Value		87%			89%
Accuracy		87%			85%

P-values <0.05 in bold; AUC = area under curve; OR = odds ratio; CI = confidence interval

Table 3

Modified antenatal scoring system for predicting the risk of Emergency CS.

Age (years)		Height (cm)		BMI (kg/m2)		CES-D score		Previous vaginal delivery		Hypertensive Disorders of Pregnancy
Value	Score	Value	Score	Value	Score	Value	Score	Value	Score	Value	Score
≤ 30	0	≤ 160	4	< 18.5	0	Low (<10)	0	Absent	5	Absent	0
31–35	2	161–165	2	18.5–25	1	High (≥10)	2	Present	0	Present	2
> 35	4	> 165	0	> 25	3

Table 4

Emergency CS prediction risk scoring system.

Score	n (%)	Emergency CS (n, %)	Odds Ratio (95% CI)	Numbers Needed to Treat (NNT)
Grade 0 (0–5 points)	459 (22%)	10 (3%)	Reference	-
Grade 1 (6–7 points)	353 (16%)	24 (8%)	3.28 (1.55–6.94)	7
Grade 2 (8–9 points)	434 (20%)	52 (18%)	6.11 (3.06–12.19)	3
Grade 3 (≥10 points)	898 (42%)	213 (71%)	13.96 (7.32–26.61)	4

AUC: 0.70 (0.68–0.73)