How do CTG guidelines impact doctors’ decisions?

There’s a widespread assumption that new guidelines must be better than old ones, simply because new things are always better. Frida Ekengård and colleagues from Sweden have been challenging that belief. I have shared some of their research previously. In these previous studies, they showed the new(ish) Swedish guideline was less sensitive at detecting cord blood acidosis at birth than the older guideline.

The same team have taken another step in investigating the effect of fetal monitoring guidelines on clinical practice. In their new research (Ekengård, et al., 2023), they explored the impact of either the older (2009) or the newer (2017) Swedish guideline on obstetric residents decisions to recommend intervention. (In Sweden a resident is a doctor in specialist training.)

How the research was conducted

At the time the research was commenced, the 2017 guideline was being introduced. Residents were familiar with the 2009 guideline. In 2017, ten residents with one to four years of obstetric experience were presented with 446 CTG recordings and asked to interpret them using the 2009 guideline, and answer the question of whether they wanted to intervene in the woman’s labour on the basis of the CTG. The CTG recordings from the last 30 – 80 minutes before birth. Half of them were from neonates who were born with a cord blood pH of under 7.05 if they were born by vaginal birth or second stage caesarean section, or under 7.1 if they had been born by first stage caesarean section. The other CTG recordings were from infants with Apgar scores of 9 or 10 and a cord blood pH of more than 7.15.

In 2022, after the new 2017 guideline had been bedded down and residents were familiar with it, a different cohort of ten residents with similar levels of experience were asked to interpret the same collection of CTGs using the new guideline. In both the earlier and later cohorts, they were unaware of the neonatal outcome, but were told whether the recording was from the first or second stage of labour to assist with interpretation. Intervention was defined as performing fetal blood sampling or expediting birth.

Findings

Using the 2009 guideline, residents indicated they wanted to intervene in 85% of the labours with CTG recordings from infants who were born with acidosis, with 90% of the traces classified from this cohort classed as pathological. 47% of the traces from infants who were born without acidosis were also classified as pathological, and residents indicated a desire to intervene in 30%.

Using the 2017 guideline to interpret the same CTG recordings yielded quite different results. The proportion of CTGs from acidotic infants interpreted as pathological fell to 72% and intervention was advised for 76%. When looking at CTGs from non-acidotic infants, the CTG was considered pathological in 24% and intervention was advised for 22%. The differences between the two guidelines for the proportion classified as pathological and when intervention was advised were both statistically significant.

The study authors also looked at interobserver variability. This is a known issue with CTG interpretation and I have written about it before. In this instance, they compared how well the interpretation using the earlier guideline aligned with the interpretation when using the more recent one. The correlation was weak (a Kappa coefficient of 0.47). When looking only at the difference in recordings that were classified as pathological when using each guideline there was a difference in interpretation of 27%. There was slightly better correlation comparing the perceived need for intervention (this is the “I might not know what to call it, but I know what to do about it” interpretation argument). 80% of the CTGs identified as warranting intervention by one guideline were also identified as warranting intervention by the other. This means that in one in 5 labours, the plan to intervene differed when a different guideline was applied.

Concluding thoughts

This research demonstrates that fetal monitoring guidelines modify doctors decisions about whether to intervene in labour. Doctors using the more recent Swedish guideline were more likely to miss cases of fetal acidosis and were less likely to intervene when acidosis was present than when they used the previous guideline. The caveat here is this study asked doctors what they would do in a theoretical situation, and this might differ from what they actually do in practice. It does seem that intention matches practice however, as national auditing of outcomes following the introduction of the new guideline showed a rise in the number of neonates born with acidaemia (Jonsson et al., 2022, here’s my summary of that research).

On the other hand, doctors in this study were more likely to interpret the CTG as normal when the fetus was not acidotic and were less likely to intervene in these circumstances, avoiding unnecessary intervention. This is a positive outcome and aligns with the small reduction in caesarean and instrumental births documented by Jonsson and colleagues. As I wrote in my post about Ekengård’s previous research:

There is no such thing as a perfect guideline that manages to always correctly distinguish well babies from those who are not. Someone has to make a decision about where to draw the line. It is therefore important that people who make use of maternity services are included in a meaningful way when guidelines are developed. They are the ones who deal with the daily and ongoing consequences of the application of guidelines in clinical practice. They should have a say about what risks are acceptable and what are not, and where the lines should be drawn.

The other take home message this research reinforced for me is the importance of acknowledging context. When we make statements like – “CTG monitoring increases the caesarean section rate by 63%” – it should best be followed with a footnote saying – well it depends… It is very easy to make blanket statements that presume that maternity professionals using CTG monitoring behave the same way everywhere and at all points in time. All aspects of maternity care arise as consequences of the decisions that imperfect humans make, often structured by imperfect guidelines, derived from an imperfect or absent research base. Variability is to be expected.

References

Alfirevic, Z., Devane, D., Gyte, G. M. L., & Cuthbert, A. (2017). Continuous cardiotocography (CTG) as a form of electronic fetal monitoring (EFM) for fetal assessment during labour. Cochrane Database of Systematic Reviews, 2(CD006066), 1-137. doi:10.1002/14651858.CD006066.pub3

Ekengård, F., Cardell, M., & Herbst, A. (2023, Apr 26). CTG interpretation templates affect residents’ decision making. European Journal of Obstetrics & Gynecology and Reproductive Biology, 285, 148-152. https://doi.org/10.1016/j.ejogrb.2023.04.022

Jonsson, M., Soderling, J., Ladfors, L., Nordstrom, L., Nilsson, M., Algovik, M., Norman, M., & Holzmann, M. (2022). Implementation of a revised classification for intrapartum fetal heart rate monitoring and association to birth outcome: A national cohort study. Acta Obstetricia et Gynecologica Scandinavica, 101(2), 183-192. https://doi.org/10.1111/aogs.14296