Much of our research in healthcare aims to show whether doing, or not doing, a thing (using a medicine, performing an operation, running a test) makes something better. These somethings are called outcome measures. What outcome measures get chosen for research and what don’t has an important impact on the generation of new knowledge. Research plays a vital role in driving change in healthcare policy and provision so decisions about which outcome measures to use are not simply intellectual arguments. The decisions researchers make about their outcome measures have profound impacts on what our healthcare systems will eventually look like.
In maternity care, we have built knowledge using outcome measures that haven’t always been the most useful somethings to decide to focus on. I see these appearing in the historical (and sometimes current) research about fetal monitoring. Here are some of the examples of the problems I have seen.
Focussing more on outcomes for the fetus / baby and not for the woman giving birth
We live in a society that has not (and to a large extent still does not) value women. This is particularly obvious when we look at women’s reproductive role in society. In research, when a fetus occupies space within the body of a woman, the health outcomes of that woman are sometimes deprioritised in comparison to that of the fetus. Looking at the trials included in the 2017 Alfirevic et al. Cochrane review on intrapartum fetal heart rate monitoring methods, outcomes affecting the fetus and baby were measured more often than outcomes impacting the pregnant woman. Perinatal mortality was assessed in 33,513 participants across eleven studies and neonatal seizures in 32,386 participants across nine studies. On the other hand, caesarean section rates were assessed in 18,861 participants in eleven studies, instrumental birth rates in 18,615 participants in ten studies, and the use of pharmacological pain management in only 1,677 participants in three studies.
While some of the trials in the review included women with multiple pregnancy this doesn’t account for the vast difference in the numbers when the outcome is about the woman rather than the fetus / baby. Researchers don’t seem to have been as interested in knowing what happened to women’s bodies as they were to knowing what happened to the fetus / baby. Researchers also seem rather disinterested in women’s experiences of pain. There is ample discussion in both academic and non-academic circles of the challenges women face in having their pain taken seriously in healthcare settings, so this is not new, nor has it gone away.
Measuring short- and not long-term outcomes
When women and their family members think about what they want from fetal monitoring, it most likely boils down to having a healthy baby. Parents are likely to be more interested in long-term rather than short-term measures. A healthy child who does what healthy children do at 3 months, a year, ten years of age, is something parents want. I suspect that few parents aim to have a baby with an Apgar score of more than seven at five minutes of age.
Measuring long-term outcomes is hard work for researchers. If you want to know what happens to a child over a ten year period, you need a study that lasts for at least ten years and you need to maintain contact with all those families over that time period. That costs a lot of money – and research funding is hard to come by. That is why only two of the 13 randomised controlled trials about CTG monitoring in labour collected longer term outcome data. If we want good long-term data we need research funders to support (perhaps even require) researchers to collect long-term outcome data.
Failing to ask users of maternity care about what outcomes matter to them
Historically, researchers haven’t always asked women and those around them about what outcomes are important to them, and then followed through to design research that provides answers that women want. For instance, no data was collected in the fetal monitoring in labour RCTs about the impact that monitoring methods had on mobility in labour, levels of physical discomfort, whether the method reduced or increased anxiety, what impact the method had on women’s privacy, the number of vaginal examinations performed during labour, whether women were able to assume the birth position they preferred, the impact on their pelvic floor function, and so on. Fortunately, much work is now being done to co-develop “patient reported outcome measures” (PROMs) and “patient reported experience measures” (PREMs) to ensure that future research provides us with broader evidence about the benefits and / or harms of the interventions we seek to introduce into practice. Sadly, it is unlikely that we will be able to turn back the hands of time to revisit the CTG vs IA question to get answers to the questions that women might now pose about fetal monitoring.
Measuring decisions and calling them outcomes
You would think that counting whether women had a caesarean section, or not, would be uncomplicated, but it isn’t. Many of the outcome measures used in the fetal monitoring research reflect decisions made by people providing care for the woman during labour. (In an ideal world the outcome measures I’m talking about here would reflect women’s decisions about their care but it is clear when reading the research that this wasn’t what was happening.)
What measures were decisions rather than outcomes? Caesarean and instrumental birth (that is forceps or vacuum assisted births), the use of pharmacological analgesia including epidurals for labour, whether fetal blood sampling was used, whether the baby was admitted to the nursery and for how long, and whether oxytocin was used to augment (speed up) labour progress are all decisions that were made by someone at a particular point in time and within a specific material context.
In all the randomised trials about fetal monitoring in labour, care providers knew whether the fetus was monitored with a CTG or with IA and might make different decisions on this basis. In an environment where there is a high level of fear about being sued for having “sat on” an abnormal CTG trace, clinicians are likely to favour surgical birth. In a healthcare system that is stretched to the limit, rates of admission to the nursery are likely to fall (because there are simply no cots available). When staffing budgets are cut and there aren’t enough midwives to go around, augmenting labour to get women back out the door of the birth suite faster makes sense. Decisions involve people in complex healthcare systems and it can all get rather messy when working out what is really going on.
Being clear about which outcome measures are decisions is important. Context profoundly impacts what the outcomes of a research trial are. Great research clearly describes this context and explains how decisions were made and by whom. Critical reading of research is vital before it is translated into practice because you can’t assume that the same outcomes will be seen in a different context.
Not being really clear about definitions
Injury to the fetal brain due to low oxygen levels in labour is one of the issues that clinicians are seeking to avoid when using fetal heart rate monitoring in labour. But what exactly is brain injury? The trials included in the Alfirevic et al., 2017 Cochrane review span a thirty year period from 1976 to 2006. Our knowledge about the developing brain has changed over that time, and the language we use to describe the same phenomena has also shifted. For instance, the Melbourne based trial published in 1976 (Renou, et al., 1976) used the terms “brain damage”, “perinatal hypoxia”, and “cerebral hypoxia” as mutually exclusive outcome categories, without defining any of them in the methods section of the paper. It is impossible to know whether these are the same as hypoxic-ischaemic encephalopathy diagnosed and graded using the schema provided by Sarnat and Sarnat that is now commonly used. It is vital that researchers are clear about what they are measuring and what they think they are measuring really is what they are measuring (for example that those seizures were due to hypoxic brain damage that occurred during labour rather than a metabolic disorder).
Focussing more on the current pregnancy and birth rather than the entirety of a woman’s reproductive journey
This issue ties back into the previous one about how researchers fail focus on long-term outcomes. Many women will go on to have another pregnancy after the current one, and all continue to have experiences related to their reproductive system. There is evidence that women who have a caesarean section are more likely to experience stillbirth in their next pregnancy (Keag, et al., 2018), and we know from the Alfirevic Cochrane review (2017) that CTG use is associated with a higher rate of caesarean section. Might we be increasing the stillbirth rate across a woman’s lifespan by using CTG monitoring in her first labour? Women who have given birth by caesarean section are also more likely to encounter complications if they later have a hysterectomy (Lindquist, et al., 2017). A higher caesarean section rate secondary to CTG use might therefore contribute to gynaecological complications later in life. Really excellent fetal monitoring research would look at not just outcomes in the current pregnancy, but across the woman’s subsequent pregnancies and the remainder of her reproductive life span.
Choose your tools wisely
So in summary, dear fellow researchers, choose your outcome measures wisely in collaboration with the woman and others who will bear the consequences of the outcome(s) you are interested in understanding. And dear fellow readers of research, look for and reflect on the decisions researchers have made about what to measure and the context in which their research was conducted. As much as we might want it to, research never provides uncomplicated answers to complicated questions.
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. https://doi.org/10.1002/14651858.CD006066.pub3
Keag, O. E., Norman, J. E., & Stock, S. J. (2018, Jan 23). Long-term risks and benefits associated with cesarean delivery for mother, baby, and subsequent pregnancies: Systematic review and meta-analysis. PLoS Medicine, 15(1), e1002494-1002422. https://doi.org/10.1371/journal.pmed.1002494
Lindquist, S. A. I., Shah, N., Overgaard, C., Torp-Pedersen, C., Glavind, K., Larsen, T., Plough, A., Galvin, G., & Knudsen, A. (2017). Association of previous cesarean delivery with surgical complications after a hysterectomy later in life. JAMA surgery, 152(12), 1148-1155. https://doi.org/10.1001/jamasurg.2017.2825
Renou, P., Chang, A., Anderson, I., & Wood, C. (1976). Controlled trial of fetal intensive care. American Journal of Obstetrics and Gynecology, 126(4), 470-476.