Comparing the effect of STan (cardiotocographic electronic fetal monitoring (CTG) plus analysis of the ST segment of the fetal electrocardiogram) with CTG alone on emergency caesarean section rates: study protocol for the STan Australian Randomised controlled Trial (START)

We believe that STan monitoring (cardiotocographic electronic fetal monitoring (CTG) plus analysis of the ST segment of the fetal electrocardiogram) of labouring women will result in a reduction in the proportion of emergency caesarean sections when compared with CTG monitoring alone, from 17% to 12%. The primary hypothesis is that the proportion of emergency caesarean sections for women on STan will not be equal to that for women on CTG monitoring alone. The study will be powered to detect an absolute difference of 5% with a conservative two-sided alternative. Our secondary hypotheses are that: there will be similarity in neonatal and maternal clinical outcomes in both treatment arms; STan monitoring will result in improved psychosocial outcomes; STan monitoring will be cost-effective compared to CTG alone; and STan monitoring will be preferred by women and they will be willing to trade-off the disadvantages, such as reduced mobility and the need for a fetal scalp electrode (FSE), in order to obtain advantages such as reduced chance of caesarean section.

This study will be a single-centre, parallel-group, randomised controlled trial, conducted over 3 years. Figure 1 provides the study flow and Fig. 2 shows a version of the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) figure for the trial. Details of the study methodology are outlined in the SPIRIT checklist in Additional file 1. There will be an intervention group (CTG + STan) and a conventional treatment group (CTG).

The trial is administered by the University of Adelaide, Adelaide, South Australia and the Women’s and Children’s Hospital, Women’s and Children’s Health Network, Adelaide, South Australia. A total of 1818 women who have antepartum or intrapartum risk factors that increase the risk of fetal compromise for which continuous electronic fetal monitoring during labour is recommended will be randomised for the study. Participants will be recruited at the Women’s and Children’s Hospital, a high-risk specialty facility with approximately 5000 deliveries per annum. Women are eligible for the trial if they meet the following inclusion criteria: 18 years or older; capable of informed consent; literate in English; and singleton fetus in cephalic presentation. Women will be excluded from participating if: they are less than 36 weeks gestation; they are planning a caesarean birth or require a caesarean due to placenta previa or vasa previa; they have contraindications for use of a fetal scalp electrode; they do not require continuous electronic fetal monitoring; they have participated in the study in a previous pregnancy; or there are known fetal structural or functional cardiac conditions.

Considering the results of our pilot study [25], current hospital rates for emergency caesarean section and electronic fetal monitoring, and the results of previous research [12], we powered the primary hypothesis on a reduction of emergency caesarean section from 17% to 12%, (i.e. a 5% absolute difference), with 80% power and a conservative two-sided α = 0.05. To detect this absolute difference between the treatment groups, a total sample size of 1634 women is required. After allowing for a very conservative 10% dropout after consent and a further 22% attrition after consent due to no clinical indication for fetal monitoring in labour, and a possible further 5% non-compliance in the STan + CTG arm, consent would need to be gained from at least 2588 women. Given a recruitment period of about 2 years, we estimated that it would be possible to consent 2588 women, of whom 1818 will be subsequently randomised. This sample size would also provide power to detect clinically meaningful differences in the secondary outcomes of interest.

This is achievable at the study hospital. Of 5000 women per year birthing at the Women’s and Children’s Hospital, 12% have elective caesarean sections and do not labour, and approximately 8% are born pre-term, leaving 80% of women to labour at term (4000 women per year). Of these women, approximately 80% have electronic fetal monitoring (3200 women per year). The US study [14] reported a consent and randomisation rate of 25.6%. Assuming a conservative consent and randomisation rate of only 20% of the over 9600 women thus eligible over a 3-year period, 1920 women would be consented and randomised. Given the 1818 consented and randomised women required in our sample size calculation, this sample size should be achievable.

The randomised controlled trial was approved by the Women’s and Children’s Health Network’s Human Research Ethics Committee (HREC/17/WCHN/14) and is covered by indemnity insurance. Original consent forms will be stored in a locked filing cabinet, in a locked room. Data will be stored on secure servers at the University of Adelaide and the Women’s and Children’s Hospital. Data and consents will be retained for 15 years after the publication of the trial according to the standards required by the Australian Code for the Responsible Conduct of Research [26].

Women will be approached to ascertain interest in and to seek consent about the study at 32–36 weeks gestation. Consent will be obtained by appropriately trained clinicians and will take place at various locations throughout the hospital including the labour ward, the Women’s Assessment Service, antenatal/gynaecology wards and antenatal outpatient clinics. After ongoing assessment of eligibility during labour, eligible consenting women will be randomised once the decision to electronically monitor the fetus during labour is made. Any excluded or voluntarily withdrawn patients will receive usual care without prejudice. Demographic and clinical baseline data will be collected by the research midwife at or soon after trial entry. Clinical observation will commence at randomisation and will end 6 weeks postnatally. The last contact for most women will be at approximately 7 weeks post delivery with the receipt of a psychosocial questionnaire. A subset of women consenting to additional follow-up will be contacted for a face-to-face interview after the questionnaire has been returned and a sub-sample of consecutive women will also be asked to complete a second questionnaire at about 16 weeks post delivery to understand patient preferences using a Discrete Choice Experiment (DCE) survey (see Fig. 2).

Once consent is obtained and eligibility criteria are met, participants will be randomised to either the treatment group (CTG + STan) or the conventional group (CTG only). Simple randomisation with stratification for parity will be implemented to produce a pre-determined schedule with an allocation ratio of 1:1, with treatment allocations accessed by a telephone-based system provided by NHMRC Clinical Trials Centre at the University of Sydney.

Continuous electronic fetal monitoring will be conducted by midwives and obstetricians who have been trained in the use of CTG and STan monitoring (holding Australian national Fetal Surveillance Education Program (FSEP) CTG accreditation, as well as in-house, institutional accreditation for competency at STan use and interpretation). Within 6 weeks of birth, all labours and deliveries will be reviewed by a multidisciplinary panel of experienced clinicians to assess adherence to electronic fetal monitoring (CTG and STan) protocols and procedures. Any evidence of protocol and STan guideline violations will be fed back to the relevant providers to optimise protocol adherence.

Maternal and neonatal data will be collected from retrospective case-note review by the research midwife who is not involved in the primary provision of care. The treatment allocation cannot be concealed because the content of the notes will reveal the care received. An independent review of 20 records will be completed to validate the quality and accuracy of data entry. Outcomes and timepoints of data collection are detailed in Fig. 2. Prospectively collected data will also be obtained in order to construct a CONSORT flow chart [29, 30].

Analysis of the randomised controlled trial will be performed by the Data, Design, Statistics and Research IT Service of Adelaide Health Technology Assessment (AHTA), School of Public Health, The University of Adelaide. Statistical analysis will follow standard methods for randomised trials and will be performed on an intention-to-treat basis, according to treatment allocation at randomisation. All analyses will follow a pre-specified statistical analysis plan. For dichotomous outcomes including the primary outcome of caesarean section (yes/no), proportions will be compared between treatment groups using the relative risk with 95% CI, obtained from a log binomial regression analysis with adjustment for strata (defined by parity) used in the randomisation. Further log binomial regression analyses (for sensitivity) will examine the effect of adjustment for parity, as well as pre-specified prognostic baseline variables and/or variables not balanced (by chance) at baseline, with results reported as the relative risk with corresponding 95% CIs. Categorical outcomes with more than two categories will be modelled with multinomial or ordinal logistic regression (or an appropriate generalised linear model alternative) with adjustment for parity. Proportions will be compared between treatment groups with relative risks and 95% CIs where possible. Continuous outcomes will be compared using differences between mean values estimated from linear or generalised linear regressions (depending on the nature of the data) adjusted for parity (with 95% CIs). Secondary analyses will explore evidence for heterogeneity of effects between the two parity strata using interaction tests and subgroup analyses. All subgroup comparisons will be pre-specified and performed with appropriate tests of interaction. Although this study will be underpowered to determine equivalence of neonatal outcomes or maternal secondary outcomes, the neonatal data will be able to be incorporated into meta-analyses including neonatal data from previous clinical trials (including the use of individual patient data).

For the psychosocial survey, analysis of the data will be blinded to the specific treatment group, with simple indicators ‘A’ and ‘B’ provided by AHTA.

For the DCE, analysis will begin using multinomial logit (MNL) models and mixed logit models; alternative model specifications (e.g. latent class models) may also be used to explore preference heterogeneity. Models will be evaluated for goodness of fit using the likelihood-ratio χ² statistic for the global test of zero model coefficients, McFadden’s pseudo R² and Akaike’s information criterion (AIC). MNL and mixed logit model results will be presented as β parameters and odds ratios with 95% confidence intervals. Trade-offs between parameters (e.g. benefit–harm trade-offs) will also be estimated; and will be calculated as the ratio of model parameters.

For the cost-effectiveness study, data collected from clinical feeder systems at the hospital will be calculated on a full absorption basis including labour and postnatal ward, medical and midwifery staff, pathology, pharmacy, operating theatre, medical and surgical supplies, goods and services. Preparation of costing data will conform to Australian Hospital Patient Costing Standards [38]. Direct cost comparisons will use case-mix classifications (Australian Related Diagnosis Groups classification, version 8.0) [39]. The incremental cost per emergency caesarean section avoided will be calculated. Bootstrapping will estimate a distribution around incremental costs and incremental health outcomes, and confidence limits around the incremental cost-effectiveness ratio will be estimated. One-way sensitivity analysis will be conducted around key variables, and a probabilistic sensitivity analysis will be conducted to estimate joint uncertainty in all parameters.

Adverse events will be monitored for by the clinical investigators and research midwife at ward rounds or perinatal review sessions. The research midwife will also monitor for adverse events as she extracts clinical data from the case notes. Adverse events entered into the data management software will automatically trigger the generation of an email alert to the obstetrician and paediatrician investigators. Additionally, results of the EPDS within the psychosocial questionnaire will be monitored to ascertain offers of referral.

A Data Safety Monitoring Committee (DSMC) has been formed to assess emerging external evidence in relation to STan, and to monitor adverse events, compliance with trial protocol and progress of recruitment. The DSMC will be notified of and meet as soon as possible after the occurrence of any of the following adverse events: maternal death; and fetal or neonatal death that could be reasonably attributed to, or partly caused by, CTG + STan or CTG alone. The DSMC will also be notified of and asked to deliberate on adverse events including unplanned intensive care admissions (maternal or neonatal) and any other adverse event of significant concern to the investigators or clinical staff. The committee will also be convened in the event that the chief investigator (CW) receives a report of an adverse event, due to STan monitoring, from weekly alerts of the published literature. Responses of the DSMC may include cessation of the study or modification of the protocol, which will in turn be re-submitted to the ethics committees.

Results will be disseminated via peer-reviewed publications as well as at institutional and state-based clinical meetings and national and international conferences such as the Perinatal Society of Australia and New Zealand and the Royal College of Obstetrics and Gynaecology. It is expected that the study will also be the subject of journal clubs and Twitter feeds, and we will ensure that the results are translated into relevant clinical guidelines, practice and policy briefs, news releases, educational sessions and one-to-one briefings. Results will also be directly disseminated via the Women’s Healthcare Australasia Clinical Network.

The trial is currently recruiting participants. Current protocol version 28 (13 March 2019). Recruitment commenced in 22 January 2018; expected recruitment completion is July 2020. The primary investigator will notify the trial register, the ethics committees, the DSMC and this journal should any major changes be made to the trial protocol.