Introduction

Cardiotocography (CTG) is a commonly used investigative modality in maternity care to evaluate fetal condition; it is frequently used in pregnant women classified at higher risk of complications, such as perinatal mortality.. Over recent years, the number of high-risk obstetric patients has substantially increased, primarily attributed to recent guideline changes which have lowered diagnostic thresholds for obstetric conditions, implementation of national initiatives, combined with an ageing maternal population [1–5]. Consequently, this has placed a significant strain on the outpatient service to accommodate the increased demand for serial monitoring, such as antenatal CTG. Concurrently, frequent requirements for hospital monitoring can pose a psychological, emotional and monetary strain for women during the antenatal period [6,7]. Frequent travel to the hospital can also pose an environmental impact which is an importation consideration to mediate following the NHS plan to reach net zero by 2040 [8]. Therefore, discovering pathways to address such concerns is paramount to enhancing the quality of current obstetric provision.

Advancements in digital medical technology combined with the increased access to wireless internet has enabled the innovation of sophisticated remote CTG technology which can be conducted outside of the hospital setting [9]. The recent pandemic has also had a substantial influence on technology development, as obstetric services now need to consider alternative solutions to current healthcare provision standards [10]. Remote digital maternal-fetal monitoring represents potential advantages by improving patient autonomy and education, whilst potentially reducing hospital attendance and overall economic burden [9,11]. Importantly, remote CTG presents opportunities to improve overall access to fetal monitoring services and aims to address current health outcome disparities seen across the obstetric population [12]. Despite this, widespread implementation has not yet been achieved due to the possible safety risks and cost implications, whilst clearly the appropriate indications for routine usage within pregnancy must be determined. A single prior review on the effectiveness of remote CTG monitoring has demonstrated favourable outcomes, however, combined both remote home and remote in-hospital monitoring at different gestational periods, proving difficulty for researchers to delineate the true benefit of home monitoring [13].

Before home antenatal CTG can be routinely implemented into care practice, it is vital to determine whether it is equivalent or non-inferior to conventional care primarily with respect to clinical outcomes, such as the identification of fetal compromise. This enables clinicians to understand the current evidence and elucidate whether this monitoring strategy is beneficial and deemed worth to continue to research. Outcomes such as perinatal mortality is a pertinent outcome to assess in novel monitoring strategies and can frequently be a definitive reason to discontinue an intervention, however, it can also be troublesome to comprehensively evaluate due to its rare occurrence. Therefore, in order to adequately determine the entire effectiveness of obstetric monitoring strategies, such as home CTG, it is essential to additionally employ a global stance and examine all relevant clinical outcomes, including its impact on healthcare usage. This further encompasses the feasibility, acceptability and economic implications associated with remote CTG. In the case of home CTG, a superior participant experience and beneficial impact on service led outcomes, combined with a reduction in associated costs may be required to consider this intervention as an acceptable alternative to routine care., To the best of our knowledge, there are currently no systematic reviews collating multiple outcome data regarding home CTG monitoring. Therefore, we aimed to comprehensively examine such outcomes to formally evaluate the applicability of home antenatal CTG monitoring for the obstetric population.

Methods

Inclusion and exclusion criteria

Inclusion and exclusion criteria can be visualised within Table 1. Principally, all primary research studies on remote antenatal CTG usage within obstetrics, whereby clinical assessment was undertaken in the home setting with additional tele-transmission were included. Studies using alternative terminology aside from CTG, such as non-stress test, but the principle of fetal monitoring remained the same, were included. Remote intrapartum CTG studies were excluded. Studies using alternative electric monitoring devices, such as fetal electrocardiogram, phonocardiotocography and magnetocardiography, whereby the data was not converted to a traditional, interpretable CTG recording, were excluded. Conference abstracts, communication papers, editorials, systematic reviews, narrative reviews and single case reports were excluded.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Inclusion and exclusion criteria.

https://doi.org/10.1371/journal.pdig.0001184.t001

Results

The database and grey literature searches identified 7695 abstracts. A total of 3352 duplicate abstracts were removed, leaving 4343 abstracts for screening. A total of 175 full texts were reviewed, from which 39 papers (28 clinical observational studies, seven randomised controlled trials and four qualitative studies), comprising 7240 participants, recruiting across 13 countries were eligible for inclusion (Fig 1) [9,11,27–63]. The most common reason for exclusion during full text examination was that the usage of CTG was not conducted in the home setting. Excluded studies following full-text screening are presented in S4 Table.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flow diagram.

https://doi.org/10.1371/journal.pdig.0001184.g001

All articles were published between 1986 and 2025, with 46% of publications occurring between 2019–2025. There were no included studies from low- and middle-income countries. One study occurred in each of: Canada, Italy, Norway and Poland. Two studies in each of: France, Denmark and Isreal. Three studies in Australia. Four studies in each of: China and the United States of America. Five studies in Japan, six in the Netherlands and seven in the United Kingdom (UK). A graphical dispersion of study recruitment can be seen in Fig 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. World map graphical representation of included studies recruitment origin.

Produced at https://www.mapchart.net/world.html.

https://doi.org/10.1371/journal.pdig.0001184.g002

Included studies were highly heterogenous in design, methodology and the outcomes measured. Study sample sizes ranged from 3 to 2860 participants; however, 74% of studies included 100 participants or fewer. Remote monitoring predominantly occurred within the third trimester (n = 21), with seven studies further assessing in the late second trimester and 11 studies not disclosing the gestation of monitoring. Participants complicated with high-risk obstetric conditions were primarily evaluated in the included studies (n = 21), with six studies including both low and high-risk participants. Five studies only included low-risk participants, and seven studies provided no such information.

Observational studies were broadly categorised into three main subtypes. Firstly, feasibility studies considered the capabilities and specifications of remote CTG monitoring to transmit and receive information that was deemed clinically useful [28,32,33,36,37,39–42,44–46,48–50,54–58,61–63]. Secondly, diagnostic accuracy studies evaluated the diagnostic capabilities of remote CTG monitoring, frequently in relation to conventional CTG monitoring [52,53]. Finally, pilot studies were those conducted in a clinical setting with additional defined control groups, without randomisation [38,51].

Completion of remote CTG monitoring in the home setting was frequently undertaken by the participant, however, in a subset of included studies, the community midwife completed the monitoring [30,31,34–36,47,48,58]. Despite device education playing a fundamental role in participant competency for remote monitoring completion, only 39% of studies detailed any information of device training [9,11,28,33,39,45,46,52–54,56,57,62,63]. Review of remote CTG monitoring data was either conducted in a synchronous [9,33,37,42,50,51,53,56,57,62,63] or asynchronous format [27,33,35,39,43,45,47,48,53]. However, frequently this was not clearly reported by studies [11,27,29–31,35,37,38,41,43,45,47,52,55,59–61]. Transmission of remote monitoring data was either by internet or telephone lines, with the majority of older included studies, transmitting with the latter modality. A detailed description of the study characteristics can be found in S5 Table.

Randomised controlled trials and GRADE assessment

Seven randomised controlled trials comparing home CTG with conventional care were included within this review, enabling pooled meta-analysis on 16 maternal, perinatal and healthcare outcomes [9,30,34,35,47,59,60]. Frequency of remote monitoring varied from up to four times per day [58], to daily [9,30,47], to twice weekly [60], or variable frequency based on the participants antenatal condition and needs [34,35]. When documented, conventional care comprised of either outpatient CTG monitoring [58,59] or inpatient admission [9,30,35,47]. Home CTG monitoring was either performed by participants [9,60], midwives [34,35], or both [30,47]. Birnie et al. and Monincx et al. were separate publications but noted to contain the same study data. Both studies were included in this review as particular outcome data was available exclusively in each paper. However, both studies were not included in the same meta-analysis in order to avoid duplication of data. The summary of findings including the GRADE assessment is presented in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Summary of findings table. Effect estimates and GRADE assessment of randomised controlled trials examining maternal and perinatal outcomes for remote CTG monitoring versus to standard care.

https://doi.org/10.1371/journal.pdig.0001184.t002

Overall, the certainty of evidence by GRADE assessment was either low or very low for all meta-analysed outcomes. Several studies were of high risk of bias using the Cochrane RoB 2 tool. Despite all but one study assessed remote monitoring in high-risk participants, there were variations in study design and methodology as noted above, which influenced the overall directness and subsequent impact on grading. Inconsistency and imprecision were moderate in several outcomes with variable point estimates and wide confidence intervals noted. Furthermore, heterogeneity was substantially high in a subset of outcomes, however with few studies included the validity of the presented heterogeneity is limited.

Regarding the primary maternal outcome, calculation of the absolute risk for emergency caesarean section demonstrated an increase of 21 events per 1000 women for those randomised to remote monitoring versus conventional care. However, results were non-significant (RR 1.08 [95% CI 0.67 – 1.73], I² = 0%). Furthermore, there were no significant differences in the rate of overall caesarean section, elective caesarean section, induction of labour, spontaneous delivery or instrumental delivery between each group. Statistical heterogeneity was low within all groups (Fig 3). Two studies additionally assessed patient wellbeing using the Edinburgh Postnatal Depression Scale (EPDS) [9,60]. Meta-analysis of two studies demonstrated no statistical significance, with high heterogeneity was noted (MD -1.55 [95% CI 3.90-0.80]; I² = 94%) (Fig 4). However, with few studies included in this analysis, the validity of the presented heterogeneity is limited. Bekker et al. measured the EDPS score at four weeks postnatal whereas Zhou et al. measured at one week postnatal [9,60].Furthermore, the control group in Bekker et al. consisted of inpatient admission compared to outpatient assessment in Zhou et al.; both potential reasons for the high heterogeneity seen.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Summary risk ratios presented as forest plots for assessed maternal outcomes.

https://doi.org/10.1371/journal.pdig.0001184.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Summary mean difference presented as forest plots for the assessment of patient wellbeing using the Edinburgh Postnatal Depression Scale (EPDS).

https://doi.org/10.1371/journal.pdig.0001184.g004

Regarding the primary perinatal outcome, calculation of the absolute risk for perinatal mortality demonstrated an increase of 11 events per 1000 women for those randomised to remote monitoring versus conventional care. However, results were non-significant (RR 1.97 [95% CI 0.37 – 10.58], I² = 0%). Results were moderately affected by a single study, which on closer inspection included slightly more participants with fetal anomaly and those complicated with preterm rupture of membranes which were randomised to the remote monitoring group [9]. This may attribute to the slight increase in cases of perinatal mortality seen. Additionally, there were no significant differences in the rates of preterm birth, low Apgar scores at 1 and 5 minutes, neonatal admission, gestational age at delivery and birthweight between groups (–). Statistical heterogeneity was low within all groups except preterm birth which was moderate (I² = 66%). Bekker et al. may have had more preterm deliveries in the control group as the participants were assigned to inpatient admission which likely inadvertently led to a greater change of earlier intervention [9]. However, from the two included studies only Zhou et al. defined preterm birth, defining as delivery between 28–37 gestational weeks [60].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Summary risk ratios presented as forest plots for assessed perinatal outcomes.

https://doi.org/10.1371/journal.pdig.0001184.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Summary mean difference presented as forest plots for the assessment of gestational age delivery.

https://doi.org/10.1371/journal.pdig.0001184.g006

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Summary mean difference presented as forest plots for the assessment of birthweight.

https://doi.org/10.1371/journal.pdig.0001184.g007

Outcomes relating to healthcare usage were assessed in four of the included randomised controlled trials [30,34,35,47]. However, only two outcomes were available for meta-analysis. There was no statistically significant difference in the number of inpatient admissions per participant during the study period and the length of inpatient stay were seen in those randomised to remote monitoring versus the control group (MD -0.52 [95% CI -1.38 – 0.34] and MD -4.17 [95% CI -9.15 – 0.80] respectively) (Fig 8). However, heterogeneity was high for both outcomes (I² = 85% and I² = 92% respectively). Albeit, with few studies included in this analysis, the validity of the presented heterogeneity is limited. Regarding, hospital admission and inpatient stay, it is important to note that the control groups in both Dawson et al. and Birnie et al. were randomised to directly to immediate hospital admission, likely explaining the large disparity in results between intervention and control groups in the respective outcomes [30].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Summary mean difference presented as forest plots for the assessment of healthcare usage outcomes.

https://doi.org/10.1371/journal.pdig.0001184.g008

Additionally, as expected with remote monitoring, Dawson et al. noted a significant increase in community midwifery visits (3.7 vs 1.4 visits; p = 0.002) and longer visits (33.5 vs 12.8 minutes per visit; p < 0.001). However, the number of antenatal outpatient visits was not significantly different (2.4 vs 3.2 visits; p = 0.11). Finally, Monincx et al. demonstrated a non-significant difference in the duration of postnatal admission between home monitoring and control groups (3.2 vs 3.5 days; p = 0.53). Importantly, no included randomised controlled trials clearly assessed the impact of home CTG monitoring on emergency attendance to maternity triage; an important outcome when evaluating the burden of a monitoring strategy on obstetric services.

Observational studies

Feasibility.

Twenty-three studies primarily evaluated feasibility outcomes in relation to remote CTG usage. Transmission delays and failures were scarcely reported by included studies. In synchronous systems, transmission of remote data was in real-time, however, reporting of delays, albeit likely only to be a few seconds, were not provided by included studies. Synchronous transmission failures of 7% was noted by Gonen et al. whereby 34 participants completed home CTG monitoring up to three times per week which was synchronously tele-transmitted via telephone lines [39]. Amongst asynchronous systems, whereby completed CTG traces were then subsequently transmitted for review, average transmission delays ranged between 30 seconds and 33 minutes [28,36,40]. However, Dawson et al. reported delays of up to two hours in certain cases, typically occurring on weekends [36]. Transmission failures ranged from 1-16%, however, minimal studies commented on this [36,44,48,58,61]. Primarily failure of tele-transmitted CTGs were attributed to insufficiencies in technological infrastructure support (e.g., poor modem operation), remote equipment malfunction and patient handling errors.

Remote CTG quality and interpretation was assessed by several included studies, frequently through subjective assessment by individual reviewers [28,36,37,39–42,44,47,56–58,62,63]. Overall, the ability to interpret remote data was moderate to excellent, ranging from 74-100% of all monitoring episodes, with most studies citing >90% of remote CTGs were suitable for interpretation [28,36,37,40–42,44,56,57,62,63]. Decreasing gestational age at time of monitoring and increased maternal weight were both negatively correlated with remote CTG quality [54,58,61]. Signal loss was reported in a select number of studies, ranging from 4-40% [33,37,39,63]. Unsupervised versus supervised participants were noted to have greater signal loss during home monitoring [37], whilst Gonen et al. demonstrated that repeated monitoring episodes had minimal impact on reducing the rate of signal loss (21% versus 20%) [39]. Methodological criteria for what defined a remote CTG episode as interpretable varied substantially between studies and was frequently undefined by authors. This must be taken into account when considering the above results, as the subjective nature of interpretation may introduce bias into the reported results. However, when authors defined what classified a remote CTG episode as interpretable, this widely varied, from just 40% or 70% of the trace being successfully tele-transmitted for review by Gonen et al. and Uzan et al. respectively [39,58], or requiring 10 minutes of continuous CTG monitoring by Tamaru et al [57]. It is therefore imperative for future studies to have uniform methodological criteria for remote CTG interpretation, as this enables researchers to robustly determine the suitability of a particular home device for antenatal care usage.

The feasibility of remote fetal heartrate monitoring was further assessed using either the validated system usability scale (SUS) in three studies [42,52,53] or the telehealth usability questionnaire (TUQ) in one study [62]. Total SUS and TUQ scores ranged from 76.5-88.1 out of 100 and 6.6 out of 7, indicating good to excellent participant usability. When measured against in-clinic use of the device, scores were comparable to home use [52].

The clinical utility of home CTG monitoring was further investigated within a select number of feasibility studies. Home fetal monitoring enabled early recognition and prompt intervention of fetal concerns within two studies [36,46]. Additionally, Hamm et al. evaluated a synchronous remote monitoring device, incorporating internet tele-transmission in 34 high-risk participants. Interpretability was high (93.9%), which subsequently enabled a reduction in outpatient office visits by 88.5% [42]. Similarly, Moore et al. demonstrated a reduction in hospital admissions by 53% following the introduction of a remote monitoring system for 100 participants [48].Finally, Axelrod et al. noted a statistically significant reduction in total clinic time for participants undergoing remote monitoring versus in-clinic monitoring (median 59.0 min vs. 159.0 min, P < 0.001) [62]. This was assessed in 20 women complicated with gestational diabetes whom used remote CTG in combination with urine analysis, glucose testing and ultrasound examination at home, followed by a telemedicine consultation [62].

Diagnostic accuracy.

Just two studies from the same authors primarily evaluated outcomes in relation to the diagnostic accuracy of remote CTG monitoring [52,53]. The reference standard was conventional outpatient CTG analysis, with different recruitment periods noted within each study. Both studies utilised a handheld, lightweight device, employing ultra-wide beam Doppler technology, enabling measurement of the fetal heartrate. The device further integrates a dedicated optical sensor to directly monitor the maternal heartrate from the abdomen, eliminating fetal-maternal heartrate cross-talk. The system includes a smartphone-based interface that displays the fetal heartrate trace and is calculated on a Bluetooth-connected smartphone, which is then uploaded to a clinical management system at the respective clinic. Median/average length of remote monitoring ranged from 4.6 to 28.5 minutes in both studies [52,53].

When comparing individual data points of remote CTG monitoring versus outpatient monitoring in both studies, the mean beat per minute (bpm) difference ranged from -0.3 to 0.63 (95% CI ranging from -4.89 to 5.32), demonstrating superior coincidence [52,53]. Interestingly, participants using an additional belt to help secure the device during monitoring, achieved slightly lower fetal heart rate agreement. However, mean signal loss during remote monitoring was minimal at 6.7 ± 5.9% [53]. Intraclass correlation was excellent ranging from 0.966-0.99 in both studies, demonstrating remote CTG as a promising tool for supplementing current outpatient CTG monitoring from a diagnostic accuracy perspective. Linear regression analysis of the 2022 results from Porter et al. demonstrated no evidence of proportional bias [53].

Pilot study.

The usage of remote CTG within the context of a non-randomised pilot study, whereby the control group was standard care, was demonstrated in two studies. These studies used patient-operated mobile CTG devices with tele-transmitted data via internet connectivity [38,51]. Regarding maternal outcomes, there were no significant differences in the rates of caesarean section, preterm birth or low neonatal birthweight. Regarding neonatal outcomes, Pan et al. demonstrated a significantly reduced rate of neonatal asphyxia in those under remote monitoring versus conventional care (1.58% versus 3.18%, p = 0.021) [51], however, this was reported as not significant by Gan et al (4% versus 7%, p = 0.483) [38]. Conversely, Gan et al. noted an increased rate of neonatal intensive care admission in the remote monitoring group (12% versus 8.3%, p = 0.044), however, their composite adverse neonatal outcome, comprising of several adverse perinatal sequalae (neonatal death, fetal distress, admission to NICU) was non-significant (p = 0.319) [38]. Interestingly, several demographic characteristics were evaluated by Gan et al., noting participants recruited to undertake remote CTG monitoring were more likely to be nulliparous (58.0% versus 44.6%, p < 0.001), classify as a high-risk pregnancy (68.8% versus 60.0%, p = 0.003) and to have a higher education level (57% versus 52% undergraduate education, p < 0.001). The reasoning for these disparities is unclear, albeit the authors suggest a greater proportion of nulliparous participants undertook remote monitoring as they may have a reduced sense of pregnancy fear compared with multiparous participants [38]. Furthermore, education level is closely linked to digital literacy and is a perceived barrier to telehealth update, which may explain why women with lower education levels were less prevalent in such studies [64].

Responses from participants noted that those assigned to the remote monitoring group demonstrated a more comprehensive knowledge of fetal heart rate monitoring, including appropriate interval timing, whilst further highlighting a significant reduction in anxiety scores using the self-rating anxiety score (SAS) [38,51]. Despite over 3900 participants recruited between both studies, the evaluation of system-based outcomes was minimal to obsolete. Soley, Pan et al. noted a slight reduction in hospital attendances in the remote monitoring group, however, this was not significant (7.86 versus 8.13, p = 0.283) [51]. Further prospective studies are encouraged to evaluate system-based outcomes alongside clinical outcomes, for a more comprehensive analysis of remote telemonitoring interventions.

Device safety.

When considering use of a technology with the potential to act as an adjunct or supplement to traditional antenatal care, it is vital to consider the safety provisions employed to manage device related concerns. Unfortunately, only 28% of included studies provided such information [9,36,39,41,45,54-56,58,60,62]. Brief statements were provided in five studies, whereby if remote CTG’s demonstrated abnormal findings, then attendance to hospital was advised [9,36,54,58,62]. Six studies supplied additional information referring to specific CTG recording abnormalities which would classify a remote monitoring episode as abnormal. However, this remained variable in reporting studies, ranging from just the absence of accelerations [39,45], to recurrent variable decelerations and tachycardia [56], or a comprehensive list of possible abnormal features relating to each assessed aspect of the CTG recording (baseline heartrate, variability, decelerations, accelerations) [41,55,60]. However, just two studies provided sufficient information regarding what the hospital assessment entailed, detailing either a repeat outpatient CTG and/or a sonographic assessment of the fetus [55,60]. Consequently, it is vital for future studies to improve on the adequacy of methodological detail relating to remote CTG safety concerns, ensuring researchers are able to sufficiently replicate available studies.

The median number of remote CTG’s which demonstrated abnormal findings was 6.5% (range 0–36%), albeit, 80% of studies had rates of 8% or lower [36–38,42,45,48,50,54–56]. It should be noted that the single study with 36% abnormal remote traces, only included participants with isolated structural abnormalities, thus the rate of abnormal CTGs is likely to be far greater. Frequently, abnormal monitoring episodes were repeated at home or in hospital and were subsequently normal. Adverse outcomes relating directly to the usage and the technical aspect of the equipment were rarely explored in the included studies. However, when noted, no adverse outcomes occurred in relation to research involvement [62].

Patient and provider experience.

Four of the 37 aforementioned studies primarily evaluated patient and/or provider experiences with remote CTG monitoring, either through focus group or a semi-structured interview format [11,27,29,43]. Patients frequently favoured home monitoring over hospital monitoring, often due to a reduction in travel time, increased convenience and the positive effect this had on family life [11,43]. Patients acknowledged how home monitoring can extend the patient-clinician relationship, whilst enabling patients to take an active role as their own care provider during pregnancy [27,29]. Whilst empowering, participants also recognised the challenges associated with remote monitoring, frequently attributed to the rise in clinical responsibility for one’s health [29]. Additional studies examined patient satisfaction using either a Likert scale [34,41,45,57], or yes/no answering [54,63], demonstrating high user satisfaction. Furthermore, participants cited that remote monitoring devices were easy to use and provided reassurance [44,48,62,63].

Home monitoring was perceived by stakeholders as a more comfortable and less stressful environment for patients, whereby these positive attributes would be most beneficial for certain high-risk groups requiring regular monitoring (hypertensive disease in pregnancy) [27]. Clinicians further cited benefits of home monitoring in relation to improving patient control and understanding of their own pregnancy. However, certain obstetricians and midwives did express fear of losing overall control over the patient’s pregnancy management, recommending that a wider change in clinical perspective may be required to enable the synergy between home and hospital care [27]. Furthermore, stakeholders highlighted patient safety concerns regarding remote monitoring, but emphasised that in-depth patient training, regular patient-provider communication and synchronous CTG reviews could help manage these concerns. Irrespective, it was acknowledged that patients would likely feel a greater sense of safety at home, which is an important factor to consider within the global assessment of home monitoring [27].

Economic burden.

Required capital for implementation and continued costings for maintaining remote CTG services are vital considerations. No included studies were primarily economic evaluation studies, however, several observational studies and randomised controlled trials described or analysed cost outcomes alongside clinical outcomes [9,30,32–34,39,41,48,49]. Initial startup costs were variable, but ranged from £1524 for a single remote device to £52,000 for a multisite set up, which when adjusted for inflation to today’s current cost, would be approximately £98000 [34,39,41,48]. However, universally, included studies demonstrated consistent antenatal non-fixed savings, ranging from £1552 to £2113 per participant within the remote monitoring groups, enabling swift recovery of initial implementation costs [9,30]. Frequently savings were as a result of fewer antenatal admissions [9,30,32–34], in addition to a consistent reduction in the cost and number of telemedicine encounters versus regular outpatient appointments [34,49]. Whilst Dawson et al. did note that community midwifery time and respective travel costs may increase for participants undertaking remote monitoring, there was also a reduction in time off work for both participants and partners who were randomised to domiciliary care [34]. When costs were collated with antenatal clinic visits, inpatient admission days, alongside the initial infrastructure costs, remote CTG monitoring was £223.83 cheaper than conventional care per participant for antenatal management. Therefore, economically, remote CTG monitoring may be a viable option for antenatal care, however, it is important that the initial startup costs are mediated by ensuring the service is offered to a wide enough number of women and for an extended duration to maintain viability.

Risk of bias.

The detailed results of the risk of bias assessments for all included articles are presented in S6 Table. A summary of the risk of bias results for the seven randomised controlled trials using the Cochrane RoB 2 tool can be seen in Fig 9. The overall quality of studies was either moderate (three studies) [9,34,47] or high risk of bias (four studies) [30,35,59,60]. This was primarily attributed to both participants and researchers’ awareness of the assigned intervention during the trial, which is an inherent bias due to the technology being assessed. However, the majority of studies did not provide adequate information to determine if an appropriate analysis was used to estimate the effect of assignment to intervention, thus determining a higher risk of bias. Furthermore, studies frequently did not include a pre-specified analysis plan prior to the production of results, leading to some concerns of bias surrounding the selection of reported results. Summary results for the 28 observational studies using the modified QUADAS-2 checklist are presented in Fig 10. The overall quality of the articles was heterogeneous, with moderate to high risk of bias. This was primarily attributed to the suboptimal quality of methodological reporting with respect to participant inclusion/exclusion criteria, inadequate reasoning for sample size and insufficient detail to enable study replication. Many studies lacked a reference group or comparison, restricting researchers to determine quantitative comparisons to standard care. Most commonly, included observational studies failed to provide information regarding cost implications. When provided, only brief cost statements were usually provided [32,33,39,41,48,49,53], with the exception of one study providing slightly more detailed crude cost estimations [61]. Cost assessment was more robust within included randomised controlled trials. However, it should be noted that overall risk of bias was generally lower in more recent versus older published studies. Sole qualitative studies were generally of lower bias risk across all categories assessed within the JBI frameworks, however, frequently there was unclear or inadequate information provided relating to the theoretical and cultural standpoint of the researcher.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Risk of bias results for randomised controlled trials using the Cochrane risk of bias (RoB 2) tool.

https://doi.org/10.1371/journal.pdig.0001184.g009

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 10. Risk of bias results for clinical observational studies using the QUADAS-2 checklist.

https://doi.org/10.1371/journal.pdig.0001184.g010

A summary of study findings in this review, including recommendations can be viewed in Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Summary of existing evidence.

https://doi.org/10.1371/journal.pdig.0001184.t003

Discussion

This systematic review and meta-analysis was designed to comprehensively assess the role of home CTG monitoring for obstetric care, by evaluating its impact on clinical outcomes, including assessing the feasibility, diagnostic accuracy, acceptability, utility and economic burden. Despite included studies demonstrating a high risk of bias, primarily due to methodological heterogeneity, remote CTG monitoring demonstrated non-inferiority across all maternal and perinatal outcomes. Synchronous and asynchronous remote CTG monitoring systems were feasible in a wide range of settings and the ability to interpret remote data was graded as moderate to excellent in reporting studies. However, frequently reviewer interpretation was subjective, with a variety of different definitions for when a CTG was classified as interpretable, which highlights a potential concern for bias. Transmission failures were generally low, but when occurring were frequently in relation to infrastructure and equipment errors. Transmission delays were infrequently reported, albeit more commonly cited in asynchronous systems, with periods of extended delay. Transmission delay is an important aspect to consider when minimising safety concerns relating to remote monitoring, particularly in high-risk patient groups whereby emergency intervention following fetal monitoring may infrequently occur. Therefore, remote CTG devices capable of synchronous (live interpretation) rather than asynchronous tele-transmission may be more suitable to assess in future research. The diagnostic accuracy of remote CTG monitoring was promising, with comparative capabilities to conventional outpatient CTG with respect to coincidence and beat-to-beat variability.

Overall acceptability ratings for remote CTG monitoring were high for both patient and providers, citing common benefits in relation to satisfaction, reduced travel, economic savings, increased access to obstetric care and balance in healthcare equity. Whilst implementation and infrastructure costs for remote CTG monitoring were high, costs could normally be accrued back over time due to non-fixed savings, such as consultation and travel costs from the remote service versus routine care. Methodological reporting of device safety was infrequently reported, with minimal information on specific CTG criteria warranting attendance to hospital or subsequent detail about hospital assessment. However, generally the rate of abnormal remote CTGs was less than 10% and adverse events relating to the technological capabilities of the devices was minimal. Additionally, information regarding service-led outcomes such as hospital attendance/admission and the impact on clinical workflow were scantly reported, however, are vital outcomes to assess. These outcomes are essential for evaluating the real-world burden and benefit of remote CTG. There was an underrepresentation of pilot and high quality randomised controlled trials, suggesting a need for further robust research. Furthermore, current included literature provided limited reporting of clear methodological, safety pathways and technological capabilities of the remote CTG systems, proving difficulty for researchers to accurately replicate studies. Current evidence is graded as low or very low using the GRADE approach, suggesting future studies should consider a more rigorous approach to study design, in order to develop a greater cohesive and applicable body of evidence to advocate for routine implementation in standard antenatal care pathways.

The meta-analysis findings of this review align similarly to the only prior published review by Li et al., which explored the effectiveness of remote CTG monitoring in nine randomised controlled trials [13]. Maternal and perinatal outcomes were not significantly different between remote and usual care groups. However, Li et al. additionally demonstrated a statistically significantly reduced rate (34%) of neonatal asphyxia, defined by low Apgar scores (<7), in the remote care group. Upon closer inspection of this review, it can be noted that included studies evaluated remote CTG monitoring in a range of settings and pregnancy stages. This included combining remote antenatal assessment at home alongside intrapartum examination in hospital [13]. Amalgamation of studies for meta-analysis from different pregnancy stages warrants some concern, as it poses difficultly to clearly establish which area remote CTG monitoring is most beneficial. Particularly given the wide variation of potential management strategies in the antenatal and intrapartum stages. In our updated review, which specifically focusses on home antenatal monitoring, we demonstrated a non-significant difference for neonatal asphyxia. Similarly to our review, a lack of high-level evidence was available for inclusion highlighting a clear need for an in-depth evaluation of remote monitoring on clinical maternal-fetal outcomes.

As previously noted, many included studies contained a small number of participants, which can be typical when assessing digital monitoring interventions [65]. However, in the context of a randomised controlled trial this can be troublesome for sufficiently evaluating clinical outcomes resulting in severe consequences, such as perinatal mortality. Therefore, employing a compositive adverse outcome score may prove beneficial in future research to evaluate the effectiveness of home CTG. Only, one included trial had utilised this, incorporating several outcomes such as perinatal mortality, low Apgar score, admission to NICU and caesarean section [9]. Presently, there is a lack of clear consensus regarding which clinical and service-led outcomes should be integrated into a composite outcome. The development of a core outcome set for remote obstetric studies may prove beneficial to reduce any inconsistent outcome reporting and improve the evidence for remote CTG research.

Aside from remote CTG monitoring, prior reviews have demonstrated the effectiveness and feasibility of several other obstetric telehealth interventions and remote/wearable monitoring devices, taking into consideration a range of clinical and service-led outcomes [66–69]. Consequently, in an era of heightened technology development, the diverse availability of digital healthcare interventions is promising and promotes an innovative landscape in obstetric healthcare. This is pertinent area to explore given the ever-growing capacity concerns within obstetric services, secondary to an ageing maternal population and increased classification of high-risk pregnancies which warrant serial antenatal monitoring, including CTGs [1,70]. Remote CTG monitoring, in conjunction with additional monitoring strategies, such as blood pressure, maternal heartrate, glucose monitoring and maternal temperature may be a potential avenue to modify current antenatal care models and improve the streamlining of outpatient services. In cases where high-risk patients are required to attend the hospital multiple times a month for CTG monitoring, such as in preterm prelabour rupture of membranes, the incorporation of a home-hospital hybrid antenatal CTG model may provide relief on maternity outpatient services. However, further evaluation is still required, particularly regarding the impact remote monitoring has on healthcare usage, such as unexpected inpatient admission and emergency outpatient attendance, as well as cost effectiveness. Given the research difficulty in assessing the beneficial impact of home CTG on reducing perinatal mortality, as is already seen in the case of conventional CTG, home CTG is likely to need to demonstrate superiority against these healthcare usage and economic outcomes, including acceptability compared with conventional CTG to even consider implementation. Technological aspects such as data security are equally as important to evaluate, which were significantly under-evaluated in included studies and have also been highlighted as a research gap in prior tele-health reviews [69]. Furthermore, a clear understanding of the training requirements needed for women to adequately perform the monitoring requires further evaluation and is not currently clear in the available literature. Finally, the implementation of a new antenatal technology should aim not to overburden, but rather support healthcare professionals, in what is an already busy clinical environment. Clear support networks for technology issues, safety pathways, improvements patient flow and reduced digital complexity can all aid in improved clinician satisfaction and acceptance. Finally, choosing which high-risk patient groups to undertake remote CTG monitoring versus conventional care is still unclear. The emergence of predictive models for stratifying risk of poor outcomes, such as placental growth factor testing for pre-eclampsia, may be of further use in informing which patient groups may benefit from only in-person care versus groups eligible for home care [71]. Predictive tools may assist in the management of safety concerns with remote CTG monitoring, by stratifying the risk of adverse maternal-fetal outcome. Enhancement to antenatal healthcare models is currently a recommendation for research by the National Institute for Health and Care Excellence (NICE), suggesting a national focus on developing UK healthcare services digitally, whereby remote CTG may be an avenue [72].

An important aspect to consider is the close relationship between technological advancement and the economic state of the respective country [68]. Whilst remote monitoring may improve patient access to care, CTG equipment in particular can be expensive for the healthcare provider, whilst potentially requiring the patient to have established home infrastructure (e.g., Wi-Fi connectivity) in order to use the device. This is evidenced as included studies in this review solely derive from high-income counties. Crucially, these factors, amongst many others, such as availability of electricity and costs of device maintance, can be barriers for implementation in low- and middle-income countries [73]. However, given the significantly higher rates of perinatal mortality in these countries, the need for remote fetal monitoring may be more urgent and thus future research in this area is strongly encouraged [13,74]. The develop of a low-cost remote device, capable of tele-transmission over low bandwidths may aid to reduce the geographical and global disparities seen in obstetric antenatal care.

Crucially, the endorsement of remote monitoring devices, such as CTG, stem from clinician attitudes regarding implementation. Initial insights from this review highlight generally high provider acceptability, suggesting a potential turning point in remote fetal monitoring adoption [10]. However, concerns relating to the safety of remote monitoring were noted, ensuring participants were fully educated on device usage. An important consideration regarding remote monitoring implementation is participant digital literacy. Whilst remote CTG monitoring may improve access to care, it is well-known that digital literacy is lower in those from most deprived groups [64]. Unfortunately, participants with low educational levels only made up between 4–8% of the remote monitoring cohorts in the included pilot and randomised controlled trials. Consequently, it is imperative that future research supports inclusivity of all patient groups, ensuring robust training strategies to enable a comprehensive understanding of device usage and thus maximise digital literacy.

Finally, the environmental impact of home digital technologies, such as CTG need to be considered when implementation occurs on a wider scale. Direct benefits of home CTG may include a reduction in travel and potential reduced hospital attendance, thus limiting patient transportation emissions and resource use in hospital respectively. However, one must also consider the environmental impact of device manufacturing such as the use of metals (e.g., lithium), plastics (e.g., petroleum-derived) and the associated energy-intensive manufacturing processes. Furthermore, data transmission requires server power and network infrastructure, which on a large scale may contribute to healthcare’s growing digital carbon footprint. Ideally, a reduction of single use items will reduce the environmental impact associated with remote CTG. However, items such as gel and abdominal belts are unlikely to change from single use, similar to current in-hospital monitoring, In an era of environmental scrutiny, home monitoring devices, such as CTG, need to demonstrate environmental benefits to assist with wider implementation.

Fig 11 represents a selection of barriers and/or facilitators for remote CTG implementation, which have been stratified from the evidence consolidated in this review. Minimisation of barriers and adoption of facilitators may enable clinicians to achieve important beneficial outcomes listed at the bottom of the figure.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 11. A schematic diagram illustrating several barriers and facilitators to achieving successful remote CTG implementation.

https://doi.org/10.1371/journal.pdig.0001184.g011

The extensive search, including grey literature and subsequent evaluation of a wide range of outcomes, including meta-analysis, represent the main strengths of this review. To our knowledge, this review is currently the most comprehensive assessment of remote CTG monitoring for obstetric care. Importantly, this review enables clinicians and researchers to understand the current state of home CTG monitoring and determine how best this technology can be safely implemented into future antenatal care. This review further facilitates clinicians to consider the technological requirements for optimising remote monitoring, whilst determining the correct elements required for robust future research.

With regards to limitations, the number of randomised controlled trials was minimal and limits the data which was eligible for meta-analysis and subsequent subgroup analysis. Additionally, included studies were frequently at high-risk of bias, with insufficient methodological reporting to enable complete study replication. These factors should be considered when interpreting the presented results. Furthermore, due to the methodological heterogeneity of included observational studies, a narrative review was required to consider feasibility, acceptability, economic and service-led factors. Given that these outcomes are integral to appropriate technology implementation, it is suggested that future studies clearly delineate such factors to enable a thorough quantitative analysis of remote CTG monitoring. Additionally, many included studies were relatively old, which may reduce the generalisability of our findings. This relates particularly to the frequent usage of telephone transmission for remote CTG, which is now obsolete in most high-income countries, following the development of internet digital technology transmission. In addition, all included studies were from high-income countries, whereby available capital for technological infrastructure is likely greater, compared to low- and middle-income countries, thus limiting the global applicability of this review.

Conclusion

This systematic review and meta-analysis aimed to explore the potential effectiveness, applicability and value of remote CTG monitoring for obstetric care across a wide range of outcomes. Home CTG demonstrates noninferiority to conventional antenatal care on a wide range of clinical outcomes and represents a promising avenue for antenatal care management. However, current evidence is of low or very low quality and presently, additional high-quality evidence with sufficient methodological detail and standardised outcome assessment is required prior to making definitive recommendations. However, remote CTG monitoring has the potential to alter current antenatal care models and shape the landscape of digital healthcare innovation in obstetrics.

Supporting information