The study employed a quantitative retrospective research design, utilising a checklist to analyse patients’ past medical records – this retrospective analysis aimed to evaluate midwives’ knowledge in interpreting CTG.

The study was conducted in a secondary hospital located in the Gauteng province. The selected hospital is a referral hospital for the primary healthcare clinics, community healthcare centres, and district hospitals. Consequently, the selected hospital provides intrapartum care services to high-risk pregnant women who require close foetal monitoring using CTG.

A total of 5808 files from the selected period were accessed from the hospital’s DoH (2024) database.

A randomised sampling technique was employed to ensure this study’s findings are statistically reliable and generalisable. This method ensures that each individual in the target population has an equal chance of being selected, thereby reducing selection bias and enhancing the sample’s representativeness.

Based on standard statistical calculations and assuming a population proportion of 0.5 (which provides the most conservative estimate), a sample size of 366 participants was determined to be sufficient. This sample size allows the study to achieve a 95% confidence level with a 5% margin of error, meaning the results can be interpreted with high precision and minimal risk of error. The retrieved files were organised into a month-to-month table, and data from each month were entered into an Excel spreadsheet for random file selection (Table 1).

The archives department staff then extracted the files, and each selected file was analysed using the inclusion and motivation table. A total of366 files were selected using randomised sampling using the inclusion criteria detailed in Table 2.

The sample met the inclusion criteria (Table 1): high-risk labour, eligible records of admission, presence of CTG strips, notes from midwives or advanced midwives, and files from women who delivered between February 2023 and March 2024.

A pilot study was conducted to assess a questionnaire’s clarity and data-soliciting capability. In consultation with a statistician, a sample of files from the post-natal ward, representing 10% of the study’s sample size (40 files), was collected and compared against the instrument – this comparison aimed to evaluate midwives’ record-keeping accuracy and the researcher’s CTG interpretation knowledge. This was the chosen analysis method developed by Norisus in 2003, which uses the measurement of agreement between the subject (participant) and the judge (researcher). This method was selected because the researcher is a trained midwife specialist who can judge midwives’ knowledge. The researcher maintained the worksheet’s privacy, sending it encrypted to the statistician and supervisor. No significant modifications to the worksheet or tool were required, and the supervisor and statistician approved the commencement of data collection. As the pilot study yielded no changes to the instrument, its findings were integrated into the main study results.

Data were collected using a checklist questionnaire to assess the records made by midwives on patients’ files. The checklist comprised two sections: demographics and CTG variables as seen in Appendix 1. Demographic data included high-risk file status, professional nurse, or midwifery specialist designation, presence of CTG in the file, clarity of notes, and delivery date (Table 1). Cardiotocography variables included baseline, beat-to-beat variability, accelerations, decelerations, and overall assessment of the trace. The files were assessed for an important element of completeness of record-keeping, ascertaining that the record-keeping is accurate and signals nurses’ knowledge of CTG interpretation. Data were kept safe in the researcher’s laptop, and codes were made for the files to maintain the anonymity of the patients and midwives. The archives department assisted in removing the randomly selected files, and they only used and knew the file number. The checklist developed was used to survey every file selected to see if there was agreement between the researcher and the midwives. A table was created to show the complete analysis of the data collected (Table 3).

Data analysis involved organising, reducing, and statistically testing the information obtained during data collection (Gray & Grove 2021). The researcher organised, reduced, and captured the data, then analysed it using Excel spreadsheet software. To assess the reliability of the checklist and ensure consistent interpretation across multiple reviewers, measuring agreement was essential. Drawing on the statistical principles outlined by Norušis (2003), methods such as Cohen’s Kappa or percentage agreement were used to evaluate inter-rater reliability (IRR). These methods help determine whether different coders reviewing the same CTG documentation reached similar conclusions about midwives’ recorded interpretations.

The researcher organised, reduced, and captured the data, then analysed it using Excel spreadsheet software. To ensure consistency in observational ratings from multiple coders, an evaluation of IRR was essential (Hallgren 2012). Since the study focused on collecting data retrospectively, the checklist was modified to observe variables specific to this research. The modifications were made with the assistance of a statistician to ensure alignment with the study’s criteria. A second section was added to the checklist, transforming it into a questionnaire that the researcher completed after reviewing clinical notes and CTG tracings. A pilot study was conducted to test the clarity and relevance of the checklist, thereby supporting its validity and reliability.

The researcher signed the waiver of consent form due because of the use of the patient’s records. The ethics committee accepted the waiver, and informed consent was not obtained.

The study used clinical information about women who delivered in the hospital and the clinical staff members responsible for midwifery care, whose identities must be protected. To maintain privacy and confidentiality, the researcher generated alphanumeric codes for each clinical record file, ensuring that the patients’ identities and the midwives who attended the case remained anonymous. The research information was kept in a password-encrypted document accessible to the researcher and supervisors. The researcher used statistics to report the findings, eliminating the need to name a specific clinical record, thus protecting the identity of patients and clinical staff.

An application for full ethical approval was made to the University of Johannesburg, Faculty of Health Sciences Research Ethics Committee and ethics consent was received on 10 September 2024. The ethics approval number is REC-2988-2024. This study will adhere to the following principles: informed written consent, privacy, and confidentiality.

Under this criterion, midwives recorded reassuring heart rates (110 bpm – 160 bpm) in 56.8% of cases, non-reassuring rates (100 bpm – 109 bpm) in 2.2%, and had no records for 150 cases (41.0%). In contrast, researchers documented reassuring heart rates in 94.0% of cases, non-reassuring rates (100 bpm – 109 bpm) in 2.5%, and non-reassuring rates (161 bpm – 180 bpm) in 3.6%, with no missing records.

Midwives recorded reassuring variability (5 bpm–15 bpm) in only 0.8% of cases, non-reassuring variability (< 5 bpm for > 40 min but < 90 min) in 0.5%, and had no records for 361 cases (98.6%). Researchers documented reassuring variability in 85.0% of cases, non-reassuring variability in 13.4%, and abnormal variability (< 5 bpm for ≥ 90 min) in 1.6%, with no missing records.

Under this criterion, midwives recorded no decelerations in 27.0% of cases, non-reassuring early decelerations in 3.0%, abnormal atypical variable decelerations in 1.9%, abnormal late decelerations in 0.8%, and had no records for 246 cases (67.2%). Researchers documented no decelerations in 85.2% of cases, non-reassuring early decelerations in 3.6%, abnormal atypical variable decelerations in 9.6%, abnormal late decelerations in 0.8%, and abnormal single prolonged decelerations in 0.8%, with no missing records.

The midwives had no records for 366 cases (100.0%), whilst researchers recorded reassuring accelerations in 72.4% of cases, non-reassuring absence of accelerations in 27.0%, and had no records for 2 cases (0.5%).

Under this criterion, midwives recorded reassuring traces in 63.1% of cases, suspicious traces in 4.9%, pathological traces in 0.8%, and had no records for 114 cases (31.1%). Researchers documented reassuring traces in 72.5% of cases, suspicious traces in 13.2%, and pathological traces in 14.3%, with no missing records.

The study results demonstrate significant gaps in midwives’ recording and interpretation of the CTG. In criterion 1, the midwives’ inconsistent recording of the baseline foetal heart rate is evidenced by the disparity between their records and the researcher’s analysis. This finding suggests under-reporting of cases such as foetal bradycardia and tachycardia, which are significant in identifying the inadequacies in foetal oxygenation. Criterion 2 revealed that the record of variability was missing in 98.6% cases, suggesting that 13.4% of cases of poor variability diagnosed by the researchers were undiagnosed. The absence of deceleration records in 67.2% cases suggests that the foetal distress evidenced by late decelerations was not excluded. The absence of records and interpretations of accelerations in criterion 4 suggests a disregard for foetal physiological responses during responses to labour.

The study found that midwives were inconsistent in recording their interpretations of the CTG. Literature suggests that the clinical records must be written chronologically to provide a comprehensive picture to guide the health professional’s clinical decision-making (Lee et al. 2024). This finding suggests that midwives’ inconsistent recording may lead to an unclear clinical picture and misguided clinical decision-making to address the ongoing foetal compromise (Daydulo et al. 2020). Consequently, midwives may have had inappropriate responses to the cases of foetal distress, as they rely on complete records to make sound clinical decisions to guide their interventions (Nakao et al. 2020). The literature on the perinatal problem identification program names midwives’ inappropriate response to the presence of foetal distress as a personnel-related, avoidable factor (Tukisi 2023).

The study showed that the midwives were inconsistent in recording their interpretation of the baseline heart rate, variability, decelerations, accelerations, and overall trace assessment. The literature on CTG monitoring and interpretations suggests that the five parameters must be analysed and interpreted collectively to conclude the final diagnosis of the CTG (Chandraharan et al. 2024; Daydulo et al. 2020). In addition, Olofsson, Norén and Carlsson (2018) affirms that each criterion of the CTG is diagnostic of a specific foetal condition and has clinical significance. The baseline heart rate refers to the average foetal heart rate over 10 min, which ranges between 110 and 160 beats per minute, allowing the diagnosis of foetal tachycardia and bradycardia, which may suggest the presence of foetal distress (Zoeller & Patel 2023). Variability, on the other hand, refers to the fluctuations in the foetal heart rate, which signifies the ability of the foetal neurological system to respond to mild hypoxia induced by the uterine contractions (Crequit et al. 2022). The decelerations are the temporary and rapid drop in the foetal heart rate below the baseline, indicating a sudden cut-off in the foetal blood supply to the foetus (López-Justo et al. 2021). The accelerations refer to the temporary rapid increase in the foetal heart rate from the baseline, indicating that the foetus is not subjected to intrauterine hypoxia (Michaeli et al. 2021). López-Justo et al. (2021) affirms that baseline, variability, and decelerations are interpreted collectively to provide an overall assessment of the foetal condition. The discrepancies in the recording and interpretation of the criteria suggest that the midwives may have missed interpreting the intrauterine foetal hypoxic status (Michaeli et al. 2021). The literature suggests that intrauterine hypoxia exposes the foetus to perinatal hypoxic-ischaemic encephalopathy with detrimental effects on the foetal neurological status, which can lead to cerebral palsy (Chandraharan et al. 2024). In addition, the intrauterine foetal death and early neonatal death may result in a direct compromise of patient safety (Loussert et al. 2023).

The study’s results showed that the midwives were inconsistent in recording their interpretations of the CTG monitored during the intrapartum care. The CTG trace forms part of the intrapartum records guided by the principles of record-keeping set out in the scope of practice of the registered nurse and midwives and conditions for midwives’ practice (SANC 1993, 2022; South African Government, 2022). These regulations state that midwives should keep complete, transparent, and accurate clinical records descriptive of assessments, diagnoses, treatments, and any significant changes in the patient’s condition (SANC 2022). Therefore, the midwives’ failure to consistently record their interpretations of the CTG compromises the quality of the clinical records. This finding supports the already reported outcry about the decline in the quality of clinical record-keeping in the literature (Loussert et al. 2023).

According to the results, the midwives not only compromise the quality of clinical record-keeping. However, they are also contravening their practice regulations from SANC, which could lead to their sanctions regarding acts and omissions (SANC 2014). The midwives should record their interpretations of a specific criterion, e.g. decelerations, suggesting that the midwife did not interpret clinical information that could lead to a possible intrauterine foetal hypoxia diagnosis. Consequently, midwives could be unable to draw a relevant clinical judgement to avoid conditions such as cerebral palsy and neurological injuries, which expose the midwives to litigation (Crequit et al. 2022). Therefore, midwives’ failure to record clinical data can be used as an aggravation during such litigations.

The researcher elucidates gaps and inconsistencies in midwives’ recording of their interpretation of the CTG, which may suggest a knowledge deficit on the part of the midwives. It is an expectation for all five criteria to be consistently assessed, interpreted, and recorded. The study recommends that midwives be empowered in CTG interpretation and recording. This highlights the need for comprehensive and standardised training programmes to ensure that all midwives consistently understand CTG interpretation. Therefore, regular training and certification can help bridge this gap and improve the accuracy of data recording (Willis, Hrdina & Luiselli 2020). The inconsistencies in recording have negative implications for the patient’s safety because of misinterpretation, potentially resulting in inappropriate interventions or missed opportunities for timely action. The potential risks to the patients’ safety call for standardising the existing hospital protocols on foetal monitoring to ensure mandatory recording of all five criteria of the CTG. The patients’ safety concerns suggest a need for the midwives to be re-empowered on the ethical principles and considerations guiding their practice.

The study evaluated midwives’ knowledge of CTG interpretation, and gaps regarding midwives’ recording were identified. However, the study was quantitative, and the midwives could not qualitatively account for the reasons for missing their interpretations of the CTG record. The study did not assess the midwives’ reactions to each CTG trace and neonatal outcome, which provided direction and indication of the midwives’ knowledge and insights into the CTG interpretations. The researchers and the midwives focused on the five criteria that directly assess the foetal wellbeing and excluded the specific monitoring, analysis, and interpretation of the uterine action with direct impact on the woman in labour. The CTG has a tocodynamometer used to monitor the uterine activity to diagnose hypotonic and hypertonic uterine contractions that could threaten the safety of the woman in labour and the foetus. The study focused on the midwives’ intrapartum recording and interpretations of the CTG. It excluded obstetricians whose clinical decision-making and plans of intrapartum care and interventions are guided by the CTG interpretations. Lastly, the study was limited to a single hospital; therefore, the findings cannot be generalised to other hospital sites.

The retrospective design is practical and ethical, allowing the researchers to analyse and interpret the CTG from the clinical records whilst allowing the continuity of care. The study’s findings highlight gaps in midwives’ knowledge regarding the electronic intrapartum foetal monitoring and interpretation. These gaps may be used as a point of departure towards empowerment of the midwives to render safe intrapartum care, thus reducing the deleterious maternal and neonatal outcomes. The study was limited to a single hospital, which provides deep insights relevant to the context. Consequently, the identified gaps may be used as an audit report to guide the formulation of context-specific recommendations for quality improvement in CTG interpretation and recording amongst the midwives.