The Birth of a Problem
How the uncritical medicalisation of birth may be doing more harm than good
They say that a Caesarean section feels like someone is doing the washing up inside your abdomen. There certainly was a lot of pushing and pulling, which was disconcerting despite the numbness. But as the team of doctors repaired each of the seven layers of tissue they had cut through to retrieve my baby, the overwhelming sensation for me was one of vertigo.
I was standing at the top of a precipice, physically and mentally exhausted, looking back at the precipitous climb I’d made to get to this point. At the mountain of events, procedures, and decisions that had been made to bring a healthy baby into this world, all of which had been guided by data, but which had left me with physical and psychological trauma that would ultimately impact my ability to care for my newborn child.
And my experience was not unique. Increasingly, women are finding themselves losing control of their pregnancies and births, swept into a maelstrom of intervention that claims to have their best interests at heart, but actually ends up causing damage and distress. These are interventions that have proven, in clinical research and medical practice, to be highly successful at improving mother and child mortality. But at the same time, rates of postpartum PTSD and trauma are on the rise. As I held my newborn baby for the first time, I couldn’t help but wonder if everything I had been subjected to had been strictly necessary? If, perhaps, the scientific data had led us astray, and was now doing more harm than good?
…So begins my entry for this year’s Nine Dots Prize, a writing competition that invites innovative and thoughtful responses to a simple question, whose prize is a book deal and $100,000 to support its writing. This year’s question was: ‘Is Data Failing Us?’, and I chose to write about my personal experiences with data-driven medical science.
I didn’t win. I didn’t expect to. There were more than 600 entrants and I’m sure there are people far more qualified than me to explore the wider role of data in our lives. But I chose to enter regardless, not just for the slim chance of a life-changing prize, but also to prove to myself that I could. It was as much an exercise in commitment and self-belief, as it was a test of my research and writing skill.
Now, it seems only fair to my past self that I share the fruits of my labour with the world. And if you’ve been affected by any of the issues I discuss here, then I’d love to hear from you. Leave a comment, reach out, and check out my parenting blog at thesciencebaby.com
A New Birth
Every one of us entered this world in more or less the same way. We were conceived as the biological product of two parents, grown and carried for months inside a mother’s body, and emerged, naked and screaming. Well over 100 billion humans have been born over our species’ history, not counting the trillions with whom we share an evolutionary past, and today, just over 4 babies are born worldwide every second. Nothing is more ubiquitous, or important to our survival. Consciously and unconsciously, we learn from those billions of births, to take actions that will favour not just the survival of mother and baby, but also their physical and mental health afterwards. Archaeological evidence and global cultures reveal some of those actions, unified across time and place: women are encouraged to give birth somewhere protected from the elements and other disturbances; they assume certain positions to ease the baby’s progress into the world; and both mother and baby are cleansed following the birth, for symbolic and sanitary reasons.
But things have changed. As I attempted to make sense of my own childbirth experience, I began to realise just how much of the process has been transformed in the last half-century, to the point where I believe that, in the western world at least, our experiences of reproduction and parenthood have changed more in the last 50 years than in the preceding millionyears of human history.
It all started, as so many of these things do, with the Industrial Revolution. The transformation of the working world made everything faster and offered opportunities that simply never existed before. Feminism bubbled to the surface, and some semblance of gender equality catapulted women out of their traditional roles as mere mothers and homemakers, into workers, scholars and professionals. Procreation was no longer a woman’s sole prerogative. Indeed, today the scales have tipped in the opposite direction. Many households in western economies require two incomes in order to stay afloat. When work takes priority, family takes a backseat.
As a result, the total global fertility rate - as the average number of babies born per woman worldwide - has plummeted. Today, women in the EU are having nearly half as many children as they did in the 1960s.
And yet, this radical change can only be partly attributed to socioeconomic shifts. Ultimately, parents feel more assured in their choices to have fewer children by the medical advances that will keep them and their babies safe. Healthcare improvements through the 20th Century have slashed the child and maternal mortality rates. Throughout most of human history, every second child died, but now global child mortality is just 4.3%. Two hundred years ago, up to one in a hundred mothers died during pregnancy or childbirth, whereas today in the EU it’s just 8 in every 100,000. Clearly, whatever we’re doing these days is working, but it comes at a cost. The typical modern birth experience is more medicalised than ever.
Prior to the 20th Century, very few women gave birth in a hospital, whereas it is now an explicit expectation for ensuring quality care. The World Health Organisation estimates that 15% of pregnant women develop complications severe enough to need medical intervention to avoid disability or death, and only in healthcare settings is the full range of medical interventions available. These include the tools and procedures to induce labour, and to aid delivery, like forceps, vacuum cups, and Caesarean sections, as well as myriad pain relief medications. But while this clinical approach has undoubtedly contributed to greater survival rates of children and mothers, they are not without risks themselves. Spinal blocks can cause nerve damage and paralysis, forceps can harm an otherwise healthy baby, and all surgical procedures carry the risk of infection. Even so, medical interventions find their way into far more than just the 15% of childbirth cases that really need them. Indeed, for IVF pregnancies like my own, they are explicitly recommended, regardless of the health of mother and baby. However, it is recognised that one medical procedure can often lead to the need for others, in an escalation of intervention with a corresponding increase in risk. Even if everything goes well, the recovery from a medicalised birth is not always straightforward. C-sections constitute major abdominal surgery, after which mothers are expected to immediately care for a vulnerable newborn. And yet their use is on the rise; more than 20% of all births worldwide are performed by C-section.
It seems the standard approach these days is one of pre-emptive and preventative intervention. Clinical policymakers advocate for intervention in birth because the scientific research suggests that this is the best way of dealing with complications, even before those complications arise. But the data fall far short of capturing the entire picture. Policymakers know a lot less about reproductive health than they would have us believe.
Why Do We Know So Little?
One of the principal reasons why the details of pregnancy and childbirth are still so poorly understood, compared to many other realms of medical science, is because these experiences are predominantly the domain of women.
For millennia, the majority of cultures have sidelined and controlled women, excluding them from education, employment and scholarship. As a result, issues that affect women, such as menstruation, pregnancy, and menopause, have been all but ignored until the last century, but even then, a mainly patriarchal approach to research has been quick to dismiss women’s lived experiences. As recently as the 1940s, women would be routinely hospitalised with cases of ‘hysteria’, named for the Greek word for uterus. Hysteria was seen as a uniquely female affliction - a sign of an unstable, excessively emotional state. In fact, many of these women weren’t ill at all, but were simply exercising their independence in a male-dominated world. More tragically, some apparently ‘hysterical’ women would be suffering from real neurological disorders like epilepsy, which is just as likely to affect men as well as women, but which was routinely dismissed among the latter.
The feminist movement of the last 50 years, involving women in the research of their own experiences, has made vast improvements to our understanding of pregnancy and childbirth. The plummeting maternal and child mortality rates are testament to this. But as psychiatrist Alexandra Sacks wrote in 2017, the focus is still on the baby rather than the mother, when it is clear that the mother also needs to be healthy in body and mind in order to care for the baby. Sacks popularised the word ‘matrescence’ to describe the process of becoming a mother, just as adolescence is the process of becoming an adult. But matrescence is still largely unexplored. For instance, it wasn’t until the late 2010s that scientists realised that a mother’s brain is fundamentally rewired during pregnancy and early motherhood.Compared to hundreds of years of medical advances led by men, we have less than a decade of matrescence research under our belts.
However, even accounting for the constraints of the patriarchy, there is another good reason why the data around reproduction and childbirth are surprisingly lacking, which is linked to the way in which this kind of research must be carried out.
The Problem With Research
The ideal scientific experiment investigates a single variable while keeping everything else constant. So to test whether induction of labour is necessary for a healthy birth, you would induce a mother and observe the result, and then for the same mother during the same pregnancy you would wait for spontaneous labour, and compare the outcome. However, without resorting to time travel, it’s simply not possible to rerun an event like this. So clinical tests seek to make as close comparisons as possible, with quite widely varying degrees of success.
Randomised controlled trials take a proactive approach, comparing a trial group with a control one. But these kinds of studies are complex and expensive, and there are often ethical concerns about intentionally giving or withholding life-saving treatments.
Observational studies take pre-existing datasets and attempt to identify trends in the variable of interest, whilst identifying and accounting for all other contributing factors. However, depending on the depth and quality of data collected, and the diligence of the researchers, this is often imperfectly done.
Alternatively, case control or cohort studies start by identifying the people who have undergone a procedure or experienced an outcome you’re interested in, and then seeking out people who did not have that procedure or outcome, but who are similar in other ways. Comparing the two groups may reveal a causal factor or notable trend. Unfortunately, these kinds of analyses suffer from the same problems as observational studies, relying on researchers’ abilities to account for and eliminate differences between the groups.
A deeper dive into the numerous studies of pregnancy complications and induction risk that have so influenced clinical policy, reveals that most of them are observational or cohort studies that fail to take into account many of the other factors that could be contributing to those risks. Without accounting for those factors, we can never hope to truly understand how pregnancy complications arise and when induction might be medically called for.
As such, scientific research into childbirth is stymied by the constraints of the experiments that can feasibly be done, and the results that emerge should always be scrutinised to determine just how large a pinch of salt we should be adding to our conclusions.
For one, and most obviously, we must consider the size of the group being studied. When shopping for something online, many of us are more inclined to buy something rated 4 stars by 10,000 people, than something rated 5 stars by 10 people. The same goes for clinical studies. A sample size of hundreds of thousands is more likely to yield patterns that are reflective of the population as a whole than a group of just 100. However, we should be wary of trusting the numbers implicitly. Those large samples still have to be random and unbiased to be reflective of the real situation.
Next, we should be mindful of the strength of the relationships that are reported. Confidence intervals and significance levels are ways of quantifying how reliable a study’s conclusions really are, but for those uninitiated in statistical techniques they do little more than complicate and obfuscate. In the early 20th Century, a notable theoretical physicist criticised speculative data analysis, saying that: “if you need to use statistics to prove you are right, then you are probably wrong”. Even 100 years later, the same holds true. The studies on which critical, life-saving medical decisions are based should have patterns that stick out like a sore thumb. All too often, we see ‘small but significant’ differences that are propped up by their confidence intervals, and whose relationships have every chance of being explained by sample error, systematic error, or some unaccounted demographic factor. And yet, policymakers take these marginal trends at face value, erring on the side of caution and advocating for what they see as the conservative approach. But when unnecessarily medicalised births end up causing lingering physical and mental health issues, this conservatism smacks of laziness.
How Do We Fix It?
How can we find a way out of the situation that we have made for ourselves? Recognising that there needs to be a change at all is the first fundamental step. Already, with the growing research into matrescence, and better equity for women’s health in the medical field, there are new studies being done and more phenomena being revealed. But the current approach to research is still inadequate, when it’s this very research that is driving life-changing clinical decisions. Perhaps there are lessons to be learned from other analytical fields?
The digital age has seen an exponential increase in the amount of data that’s generated and stored, and with the corresponding rise in computing power, analysts are turning to ways of harnessing this new, so-called Big Data. Everywhere that data is generated, there is the potential to use that information to understand and inform our behaviour in those fields. And matrescence, pregnancy and childbirth are no exception. Since the vast majority of pregnancies are medically supervised, with births taking place in hospitals or clinical centres, there is a huge amount of data already being collected. Midwifery notes, hospital records, even minute-by-minute vitals throughout labour and birth are all added to a complex and multifarious data pool that, for now, is sitting largely unused. In addition to this, fertility treatments such as IVF are some of the most thoroughly documented medical procedures of all, and, if properly integrated, provide a continuous record of each woman’s journey from pre-conception to birth and beyond.
Big Data would go a long way towards addressing concerns about sample size, and with such a wide range of data available, many different factors could be investigated independently or together. With the right analysis, it could be possible to identify the potential causes of pregnancy complications long before they happen, which in turn could inform a more granular clinical policy that avoids unnecessary birth interventions.
Right now, however, Big Data’s potential far exceeds its actual utility. The problem lies in figuring out how to process such large and interrelated datasets. Not every data point is created equal, and it can be hard to foresee the kind of analysis that will be necessary to tease out relevant trends. Traditionally, Big Data analysis has fallen to human statisticians, but to get the best out of these staggeringly complex datasets, we will need to turn to analytical architects that are similarly complex. In other words, we need AI.
Artificial intelligence has experienced a meteoric rise over the last few years, and it seems to offer near limitless capability for information assimilation and processing if - and it is a big if - we know how to ask the right questions. But AI is in its infancy, and we, as its mentors, have a responsibility to train it, guide it, and recognise its limitations. Right now, despite its infiltration into every aspect of our daily lives, both man and machine still have a lot to learn. Humans can’t hand over the reins to data analysis just yet, and until our digital analysts can figure out how to draw hands, or to count the R’s in the word strawberry, it’s unlikely any clinical professionals will be trusting them with life-saving medical decisions any time soon.
For now, we must make do with the data and analysis that we already have, flawed though it may be. Reproductive science has come so far in the last half-century, and there is no shortage of research guiding how new treatments and interventions can and should be safely used. But in rushing to prove that we can medicalise the process of birth, few have stopped to question if we should. It certainly saves lives, but administered arbitrarily, it can also ruin them as well.
However you look at it, the buck stops with us. It is researchers’ responsibility to consider the issues facing women during pregnancy, childbirth, and beyond, and design the best possible experiments to investigate them. It is analysts’ responsibility to consider every potential contributing factor when presenting trends and relationships. It is policymakers’ responsibility to consider all of the data available, as well as the strengths of the conclusions, when setting their clinical guidance. It is individual doctors, nurses, and midwives’ responsibilities to consider each woman’s unique situation, and apply the guidance in the most sympathetic way. And above all, it is the responsibility of all of us to advocate for a better system that protects and serves our mothers, sisters and daughters.
Ultimately, it’s not the data that are failing us, but rather us - as the analysts and the architects of our own destiny, who are failing to rise to its challenge, realise its potential, and act in a way that not only does good, but also avoids unnecessary harm as well.