Unintended consequences of common-sense choices: Maternal heart rate monitoring on the CTG

Over the past decade or so, there has been increasing recognition of the potential for confusion of the woman’s heart rate with that of her fetus during CTG monitoring, particularly in labour. The danger here is twofold. First, thinking the fetus is fine when what is being recorded is a slightly elevated maternal heart rate, but the fetal heart rate is abnormal. Second, the normal maternal heart rate looks like a fetal bradycardia, so emergency surgery is performed to achieve the birth of a baby that was never at risk.

The incidence of maternal heart rate confusion is fairly high (Oppenheimer et al., 2025), and it is recognised as a problem in legal cases. Fetal monitoring education often covers tips on how to recognise it, and most guidelines advise that some effort be applied to ensuring the fetal heart is the one that appears on the CTG. Sometimes this is achieved by continuous maternal heart rate monitoring with a pulse oximeter. I wrote about the only piece of research yet to assess the outcomes of this approach this back in 2024 (here). The results showed better outcomes, and it does seem like a common sense idea.

However, maternity care systems are incredibly complex social systems consisting of messy humans working in (often) less than ideal circumstances. Sometimes what seems logical can introduce unintended and unexpected complexities that undermine efforts to improve safety.

The research I’m looking at today was published in January 2026 (Ploran et al., 2026). The research team were curious whether adding maternal heart rate data to the CTG may place more demand on visual perception systems and ultimately reduce consistency in CTG interpretation. They had previously shown this to be the case with twin CTGs (Ploran et al., 2022). They found that showing one heart rate at a time performed better than showing both on the trace simultaneously.

How was the research done?

Obstetric nurses, physician attendants, and obstetricians with prior expertise in CTG interpretation were recruited. Thirty-two people participated, and were shown 8 – 10 minute long segments of CTG tracings from women in labour at 32 – 42 weeks gestation with a single baby. Each tracing was taken more than an hour prior to birth and included maternal heart rate tracing. Ten tracings from babies who were born with Apgar scores less than seven at five minutes were included, and another ten from babies with normal Apgar scores.

The twenty traces were then duplicated and modified to remove the maternal heart rate data. All 32 people were shown all forty traces (babies with normal Apgars, and those with low Agpars, with and without maternal heart rate information) tracings, without clinical information being provided, and an additional twelve traces were shown a second time. Participants were shown the trace for 30 seconds only and completed an assessment of the features of the trace using the NICHHD framework (commonly used in the USA). This pressure increased cognitive load and aimed to reproduce the time pressure of a busy birth suite environment.

Both inter observer variability (the degree to which different people agreed or did not agree on the features of the same CTG) and intra observer variability (the degree to which the same person agreed or not with their first assessment of the same CTG) were calculated.

What they found

First, the researchers established the baseline for intra observer agreement on the same trace. This was better for the traces from babies with normal Apgars than those with low Apgars, and better for individual features of the CTG (like whether there were decelerations or not) than for the overall assessment of the level of concern about the trace. There was notable variability in the variability – that is some people were almost always consistent in their assessment of the trace, while others had very little agreement when they reviewed the same trace a second time.

Next, they looked to see what impact the presence of the maternal heart rate data had on intra observer variability. They started with the traces from babies with normal Apgars. Looking at variability, adding the maternal heart rate led to 11% of people changing their assessment of the trace. For accelerations, this was 22%, for decelerations it was 30%, and for level of concern it was 40%.

Let that sink in: four out of ten people made a different call about how concerned they were about an identical CTG trace, simply because the mother’s heart rate was now also shown on the trace.

It was even more notable in the traces from babies with low Apgar scores – 60% of people changed at least one part of their assessment of the CTG in response to the presence of maternal heart rate information. In summary, intra observer variability was higher (less consistency) when maternal heart rate data were added.

Then they considered inter observer variability. At the baseline this was only fair for level of concern and for accelerations, and was moderate for variability and decelerations for the traces from babies with normal Apgars. It fell to low across all features for babies with low Apgars, and fell even further when maternal heart rate data were added.

What does this mean for CTG monitoring?

Adding more visual information to the CTG in the form of the maternal heart rate subconsciously impacted professional’s capacity to reliably interpret fetal heart rate patterns. The maternal heart rate patterns looked like a “reactive” fetal heart rate pattern and appeared to pull the interpretation of the CTG towards less concern, even when the fetal heart rate was not normal.

This research was done in an artificial environment, with participants sitting at a computer and being shown the traces one after the other, without interruption. This is not the same as what might happen when interpreting a CTG in a birth room or at a central monitoring station, so we don’t know if the findings translate through to different outcomes in the clinical setting.

The authors suggest that artificial intelligence interpretation of the CTG might solve the problem by replacing human interpretation (and its tendency to be pulled off course by distractions) with the consistency of an algorithm. I think there are several other potential solutions that they didn’t touch on.

Remembering that the entire point of adding the maternal heart rate was to ensure the CTG was displaying the fetal heart rate, we could:

Stop continuously recording maternal heart rate as the vast majority of that data is “noise”. Instead, staff maternity services adequately so a midwife or nurse can remain present and regularly assess the maternal heart rate in person to ensure the CTG is recording the fetus, or provide intermittent auscultation while double checking they are listening to the right heart rate
Redesign the Doppler sensor and related software in the CTG to reduce the possibility of maternal heart rate confusion
Redesign the display so maternal heart rate isn’t overlaid on the fetal heart rate trace but is still visible in a way that makes it possible to see whether there is misrecognition of the fetal heart rate
Use an AI based approach to alert when there is misrecognition of the fetal heart rate by incorporating maternal heart rate data without displaying maternal heart rate data on the trace

In essence, this is a design problem with design solutions. Redesign the way health care is funded, staffed, and provided, and / or redesign the CTG machine. I’m curious to see if it happens, or if maternity professionals are simply asked to somehow rewire their brains to solve the problem.

Confident Intermittent Auscultation: Live is on tomorrow!

The workshop is on: Thursday May 21, at 7 pm (AEST, Brisbane time, GTM+10)

This is my two-hour long, online workshop that will build your knowledge and confidence with intermittent auscultation so you can offer it to more women in your practice. One of the things I include in the workshop is a physiology-based approach to WHEN you should be listening to the fetal heart rate and for HOW LONG – so you can gather the same quality (rather than quantity) of information that others get from a CTG recording. If that interests you – register at the link below for more details. (Yes, there will be a recording available.)

Time is running short!

Enrol now

References

Oppenheimer, L., Colussi, M., Payant, L., Honey, L., Kiely, D., Ji, J., Yang, Q., MacIntyre, A., El Sheriff, R., Young, K., Woods, S., & Garber, G. (2025). The proportion of birth asphyxia associated with maternal heart rate artifact during electronic fetal monitoring in labor. Journal of Obstetrics & Gynaecology Canada, 103128. https://doi.org/10.1016/j.jogc.2025.103128

Ploran, E. J., Comfort, L., Rausch, A., Snellings, J. T., & Rochelson, B. (2026). Perceptual distortions during simultaneous continuous monitoring of fetal and maternal heart rate in laboring patients. Journal of Maternal Fetal & Neonatal Medicine, 39(1), 2614837. https://doi.org/10.1080/14767058.2026.2614837

Ploran, E. J., Soni, S., Snellings, J. T., Rausch, A., & Rochelson, B. (2022). The effect of perceptual decision-making on the interpretation of twin fetal heart rate tracings. Journal of Maternal Fetal & Neonatal Medicine, 35(11), 2116-2121. https://doi.org/10.1080/14767058.2020.1779695