Unrecognized maternal heart rate artefact in cases of perinatal mortality reported to the United States Food and Drug Administration from 2009 to 2019: a critical patient safety issue

Kiely, Daniel; Oppenheimer, Lawrence; Dornan, James

doi:10.1186/s12884-019-2660-5

Cited by 15 publications

(13 citation statements)

References 27 publications

(21 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…13 Another limitation of traditional CTG-based FHR measurements relates to the impact of MHR artifacts, which can result in potentially dangerous consequences for the fetus. 25 Although a noninvasive system to measure FHR will inherently capture maternal recordings, the large amplitude of the maternal signal ensures that it can be captured with sufficient signal-to-noise ratio to validate it and eliminate it from the raw signal. In contrast to strategies that rely on a single or a few biosensors to capture FHR and MHR signals, the use of data from multiple biosensors (8 ECG and 4 acoustic) in the Invu system allows the algorithm to remove the maternal ECG from contaminating the fetal ECG, essentially performing signal ambiguity detection, thereby reducing artifacts and the likelihood of errors in FHR calculation and interpretation.…”

Section: Original Researchmentioning

confidence: 99%

Wireless, remote solution for home fetal and maternal heart rate monitoring

Mhajna

Schwartz

Levit-Rosen

et al. 2020

American Journal of Obstetrics & Gynecology MFM

View full text Add to dashboard Cite

BACKGROUND: Access to prenatal care can be challenging due to physician shortages and rural geography. The multiple prenatal visits performed to collect basic fetal measurements lead to significant patient burden as well. The standard of care tools for fetal monitoring, external fetal heart rate monitoring with cardiotocography, as used today, must be applied by a medical professional in a healthcare setting. Novel tools to enable a remote and self-administered fetal monitoring solution would significantly alleviate some of the current barriers to care. OBJECTIVE: To compare maternal and fetal heart rate monitoring data obtained by 'Invu system' (a wireless, wearable, self-administered, fixedlocation device containing passive electrical and acoustic sensors) to cardiotocography, toward a true remote fetal monitoring solution. MATERIALS AND METHODS:A prospective, open-label, multicenter study evaluated concurrent use of Invu and cardiotocography in pregnant women, aged 18 to 50 years, with singleton pregnancies !32þ0 weeks' gestation (NCT03504189). Simultaneous recording sessions from Invu and cardiotocography lasted for !30 minutes. Data from the 8 electrical sensors and 4 acoustic sensors in the Invu belt were acquired, digitized, and sent wirelessly for analysis by an algorithm on cloud-based servers. The algorithm validates the data, preprocesses the data to remove noise, detects heartbeats independently from the two data sources (electrical and acoustic), and fuses the detected heartbeat arrays to calculate fetal heart rate (FHR) and maternal heart rate (MHR). The primary performance endpoint was Invu FHR limit of agreement within AE 10 beats per minute (bpm) of FHR measured with cardiotocography.RESULTS: A total of 147 women were included in the study analysis. The mean (SD) maternal age was 31.8 AE6.9 years, and the mean gestational age was 37.7 AE2.3 weeks. There was a highly significant correlation between FHR measurements from Invu and cardiotocography (r ¼ 0.92; P<0.0001). The 95% limits of agreement for the difference, the range within which most differences between the two measurements will lie, were -8.84 bpm to 8.24 bpm. Invu measurements of MHR were also very similar to cardiotocography and were highly significantly correlated (r ¼ 0.97; P<0.0001). No adverse events were reported during the study. CONCLUSION: Although captured by very different methods, the FHR and MHR outputs wirelessly obtained by the Invu system through passive methods were very similar to those obtained by the current standard of care. The limits of agreement for FHR measured by Invu were within a clinically acceptable AE 8 bpm of cardiotocography FHR. The Invu device uses passive technology to allow for safe, non-invasive and convenient monitoring of patients in the clinic and remotely. Further work should investigate how remote perinatal monitoring could best address some of the recent challenges seen with prenatal care and maternal and fetal outcomes.

show abstract

Section: Original Researchmentioning

confidence: 99%

Wireless, remote solution for home fetal and maternal heart rate monitoring

Mhajna

Schwartz

Levit-Rosen

et al. 2020

American Journal of Obstetrics & Gynecology MFM

View full text Add to dashboard Cite

show abstract

“…According to Reinhard et al [5], MHR periods are found in up to 90% of intrapartum recordings and account for 6.2% of the duration of the recording. This problem is particularly frequent during the second stage of delivery and can be particularly dangerous for the fetus [6,7]. For example, France's Melchior classification (used to describe the second stage of delivery [8]) suggests an erroneous type 3 FHR pattern (corresponding to bradycardia plus accelerations synchronized with uterine contractions (UCs), and for which expulsive efforts should last for less than 15 min) these cases (accounting for ≈ 4% of deliveries) correspond in fact to MHR interferences [9,10].…”

Section: Introductionmentioning

confidence: 99%

“…The software in CTG monitors sometimes also comprises an alert system based on coincidence between the MHR and FHR channels [7]. However, these coincidences have to be checked manually.…”

Section: Introductionmentioning

confidence: 99%

Use of Deep Learning to Detect the Maternal Heart Rate and False Signals on Fetal Heart Rate Recordings

Boudet¹,

l’Aulnoit²,

Peyrodie³

et al. 2022

Preprint

View full text Add to dashboard Cite

We have developed deep learning models for automatic identification of the maternal heart rate (MHR) and, more generally, false signals (FSs) on fetal heart rate (FHR) recordings. The models can be used to preprocess FHR data prior to automated analysis or as a clinical alert system to assist the practitioner. Three models were developed and used to detect (i) FSs on the MHR channel (the FSMHR model), (ii) the MHR and FSs on the Doppler FHR sensor (the FSDop model), and (iii) FSs on the scalp ECG channel (the FSScalp model). The FSDop model was the most useful because FSs are far more frequent on the Doppler FHR channel. All three models were based on a multilayer, symmetric, GRU and were trained on data recorded during the first and second stages of delivery. The FSMHR and FSDop models were also trained on antepartum recordings. The training dataset contained 1030 expert-annotated periods (mean duration: 36 min) from 635 recordings. In an initial evaluation of routine clinical practice, 30 fully annotated recordings for each sensor type (mean duration: 5 h for MHR and Doppler sensors, and 3 h for the scalp ECG sensor) were analyzed. The sensitivity, positive predictive value (PPV) and accuracy were respectively 62.20%, 87.1% and 99.90% for the FSMHR model, 93.1%, 95.6% and 99.68% for the FSDop model, and 44.6%, 87.2% and 99.93% for the FSScalp model. We built a second test dataset with a more solid ground truth by selecting 45 periods (lasting 20 min, on average) on which the Doppler FHR and scalp ECG signals were recorded simultaneously. Using scalp ECG data, the experts estimated the true FHR value more reliably and thus annotated the Doppler FHR channel more precisely. The models achieved a sensitivity of 53.3%, a PPV of 62.4%, and an accuracy of 97.29%. In comparison, two experts (blinded to the scalp ECG data) respectively achieved a sensitivity of 15.7%, a PPV of 74.3%, and an accuracy of 96.91% and a sensitivity of 60.7%, a PPV of 83.5% and an accuracy of 98.24%. Hence, the models performed at expert level (better than one expert and worse than the other), although a well-trained expert with good knowledge of FSs could probably do better in some cases. The models and datasets have been included in the Fetal Heart Rate Morphological Analysis open-source MATLAB toolbox and can be used freely for research purposes.

show abstract

“…According to Reinhard et al [5], MHR periods are found in up to 90% of intrapartum recordings and account for 6.2% of the recording. This problem is particularly frequent during the second stage of delivery and can be particularly dangerous for the fetus [6,7]. For example, France's Melchior classification used to describe the second stage of delivery [8] proposed erroneously the type 3 FHR pattern (corresponding to bradycardia plus accelerations synchronized with the uterine contractions (UC), and for which expulsive efforts should be less than 15 min) whereas all those cases (representing ≈ 4% of deliveries) are in fact FS of MHR [9,10].…”

Section: Introductionmentioning

confidence: 99%

“…The software in CTG monitors comprises also sometime an alert system based on coincidence between MHR and FHR channels [7]. However, these coincidences have to be checked manually.…”

Section: Introductionmentioning

confidence: 99%

Use of Deep Learning to Detect the Maternal Heart Rate and False Signals on Fetal Heart Rate Recordings

Boudet¹,

l’Aulnoit²,

Peyrodie³

et al. 2022

Preprint

View full text Add to dashboard Cite

We have developed deep learning models for automatic identification of the maternal heart rate (MHR) and, more generally, false signals (FSs) on fetal heart rate (FHR) recordings. The models can be used to preprocess FHR data prior to automated analysis or as a clinical alert system to assist the practitioner. Three models were developed and used to detect (i) FSs on the MHR channel (the FSMHR model), (ii) the MHR and FSs on the Doppler FHR sensor (the FSDop model), and (iii) FSs on the scalp ECG channel (the FSScalp model). The FSDop model was the most important because FSs are far more frequent on the Doppler FHR channel. All three models were based on a multilayer symmetric gated recurrent unit and were trained on data recorded during the first and second stages of delivery. The FSMHR and FSDop models were also trained on antepartum recordings. The training dataset contained 1030 expert-annotated periods (mean duration: 36 min) from 635 recordings. In an initial evaluation of routine clinical practice, 30 fully annotated recordings for each sensor type (mean duration: 5 h for MHR and Doppler sensors, and 3 h for the scalp ECG sensor) were analyzed.The sensitivity, positive predictive value (PPV) and accuracy were respectively 62.20%, 87.1% and 99.90% for the FSMHR model, 93.1%, 95.6% and 99.68% for the FSDop model, and the 44.6%, 87.2% and 99.93% for the FSScalp model. We built a second test dataset with a more solid ground truth by selecting 45 periods (lasting 20 min, on average) on which the Doppler FHR and scalp ECG signals were recorded simultaneously. Using the scalp ECG data, the experts estimate the true FHR value more reliably and thus annotated the Doppler FHR channel more precisely. The models achieved a sensitivity of 53.3%, a PPV of 62.4%, and an accuracy of 97.29%. In comparison, two experts (blinded to the scalp ECG data) achieved a sensitivity of 15.7%, a PPV of 74.3%, and an accuracy of 96.91% for expert 1 and a sensitivity of 60.7%, a PPV of 83.5% and an accuracy of 98.24% for expert 2; hence, the model performed better than one expert and worse than the other. Hence, the models performed at expert level, although a well-trained expert with good knowledge of FSs could probably do better in some cases. The models and datasets have been included in the Fetal Heart Rate Morphological Analysis open source MATLAB toolbox and can be used freely for research purposes.

show abstract

Unrecognized maternal heart rate artefact in cases of perinatal mortality reported to the United States Food and Drug Administration from 2009 to 2019: a critical patient safety issue

Cited by 15 publications

References 27 publications

Wireless, remote solution for home fetal and maternal heart rate monitoring

Wireless, remote solution for home fetal and maternal heart rate monitoring

Use of Deep Learning to Detect the Maternal Heart Rate and False Signals on Fetal Heart Rate Recordings

Use of Deep Learning to Detect the Maternal Heart Rate and False Signals on Fetal Heart Rate Recordings

Contact Info

Product

Resources

About