Induction and ethnicity: what is the evidence?

For the past few years, a series of guideline developments, audit findings, and research studies have led to a number of clinicians and professional bodies arguing that induction of labour at 39 weeks could be justifiably recommended to Black, Asian and mixed race women.

I have deep concerns about this, for a number of reasons.

There is no evidence to support the idea that induction will reduce maternal or perinatal mortality in Black and Brown families.

Proponents of offering induction often fail to consider or convey the downsides and medium- and long-term consequences. They also fail to take into account the many benefits of spontaneous labour.

And such policies tend to exacerbate the inequalities and iatrogenic risks faced by Black and Brown people.

That’s because induction of labour entails many interventions, which lead to an increase in the very kinds of situations that create the circumstances in which Black and Brown women and babies are likely to have poorer outcomes than their white counterparts.

It’s also important to think about the dangers of describing social constructs – such as race – as risk factors. As the authors of a recent letter to a medical journal eloquently pointed out, the risk factor for maternal morbidity is racism, not race.

I have written this blog post to explain some of the issues, and to point you to resources that will help you understand this area more deeply.

Evidence of inequality

But first, in case you’re new to this area, let me start by pointing you to the evidence for inequality. I have written a series of blog posts about the reports that highlight this, and they each link to evidence.

Recent evidence has highlighted the particular difficulties faced by South Asian migrant women.

You can also find the most recent MBRRACE-UK reports at https://www.npeu.ox.ac.uk/mbrrace-uk

Induction: explaining the issues

I’m also not going to repeat everything that I have written in a general sense about induction of labour on this page. I have an entire information hub on this and I have written two books on this topic for those who aren’t familiar with the issues in a general sense.

Induction and maternal ethnicity

I first wrote about induction and maternal ethnicity in 2021. Without precedent, evidence, or warning, we learned that the writers of the NICE guideline were considering adding the recommendation to offer induction at 39 weeks of pregnancy to Black and Brown women.

My post on this is here.

I was not alone in my concern.

“We are deeply concerned that if these recommendations are taken forward uncritically, they could further embed institutional racism in maternity care, strengthen racial biases and stereotypes, legitimise skin tone as clinically meaningful, pathologize healthy pregnancies in women from ethnic minority backgrounds, and undermine choice for black and brown women.” (Douglass & Lokugamage 2021).

Douglass and Lokugamage (2021) were also concerned that, “Of the studies referenced the vast majority did not record race or were unable to, or failed to, report on ethnic variation due to low numbers of minority ethnic women. This “absence” of evidence could be construed as a form of structural racism.”

No evidence of benefit

A number of us helped to effect appropriate changes to the wording of the 2021 NICE guideline on induction of labour, but an inherent tension remains in this and many other situations.

Where guidelines are diluted in ways which leave things up to the individual practitioner rather than (as is the case here) explicitly and clearly stating that there is a lack of evidence and thus no justification to offer intervention, this leaves the path open for technocratic practitioners to continue to recommend intervention without any evidence base for this. Sometimes they cite evidence (such as the study I’m going to discuss in this post) but fail to explain the study results and give proper context. In some situations, this is because they don’t understand the issues themselves.

This isn’t just happening with regard to maternal ethnicity, but to other socially constructed categories which the maternity services deem to be ‘risk factors’, such as BMI. Sadly, women who fall into more than one of these categories often report feeling undermined from all directions.

I have written a whole chapter on the evidence around induction and maternal race in my book In Your Own Time. I hope you will read it, especially if you work in or around maternity care, because it also includes information on racism and weathering. But I can summarise it succinctly by quoting one line that I wrote near the end of that chapter:

“The bottom line for those concerned about being offered induction for maternal race is that there is no evidence to suggest that it would be beneficial.” (Wickham 2021).

Getting into the data

But I also think it’s important that more people understand why some of the studies that they are told about aren’t good evidence. That’s why I explain research in reader-friendly ways in my books.

It’s also why I’m going to break down one recent study in this blog post, so that you are better equipped to have conversations with providers who may be trying to use such studies to persuade you to have intervention that won’t actually benefit you.

In 2023, researchers published an analysis of data about induction at 39 weeks, and some people used this to further their argument that Black and Brown women should be offered induction of labour.

The researchers wanted to look at whether data show that early induction might be associated with better outcomes in Black and Brown women/babies.

Overall, the study again confirmed that, “There was no significant risk difference found by ethnicity.” (Muller et al 2023).

In other words, if you are Black or Brown, there is still no evidence that induction of labour at 39 weeks reduces risk.

As is often the case in obstetrics, however, this hasn’t deterred the authors and other proponents of early induction from calling for an increased uptake of induction of labour.

So let me break down this study a bit further.

What type of study?

First, this is a cohort study, not a randomised controlled trial.

Cohort studies can help us understand about which groups are more at risk of certain things, but they’re not designed to evaluate the effectiveness of interventions. Population-level research studies also can’t tell us anything about the health or situation of individual women, babies and families. So any study of this nature is limited in what it can tell us, and we need to look at it within the context of the bigger picture.

As you know, the bigger picture in this area shows that induction doesn’t reduce adverse outcomes in Black and Brown women or their babies.

Bias, findings, and cherry picking

The second issue is that the data used in this study came from medical records, stored in maternity databases, and weren’t collected specifically for this study.

The authors acknowledge that their findings may have been affected by missing data. Researchers understand only too well that medical records are sometimes inaccurate, and that mistakes and omissions are common. The data set did not include data on maternal risk factors, congenital abnormalities in babies, or reasons for induction. All of these factors could have affected the results.

In fact, the researchers acknowledged this too. They explained that the lack of data on underlying medical problems “…may have led to an overestimation of the risk of adverse pregnancy outcomes in low-risk pregnancies.”

However, even when researchers acknowledge such things in the body of their paper, some people who want to recommend intervention ignore this and quote only the sentences in the paper which support their arguments. That’s actually the definition of ‘cherry picking.’

Why we need to read closely

The part of the paper that is being quoted by proponents of early induction says that, “women who were induced and gave birth at 39 weeks had lower risk of adverse perinatal outcomes compared to women who had expectant management and gave birth after 39 weeks.” (Muller et al 2023).

This is in ALL the women whose medical records were looked at. As a reminder, there was no risk difference when they just looked at ethnicity.

But again, it’s important to not just take one sentence out of context but to read the rest of the paper AND also consider the wider issues.

Because, as the authors acknowledge, “the difference was small and we estimate that overall 360 inductions are needed to avoid 1 adverse perinatal outcome.” (Muller et al 2023).

This is a very large number of inductions, so one thing we need to consider is what the negative effects for the other 359 women will be. As a reminder, that’s because inductions (and elective caesareans) are also associated with risks, downsides, and short, medium, and long-term effects. These are not benign alternatives, and women/families need to weigh up the pros and cons in order to make the decisions that are right for them.

While many people don’t realise it, there’s a systemic issue here as well. When lots of women are having their labour induced, the increased number of interventions mean that there are more potential adverse events, so more need for monitoring, and more caesarean sections. Additional monitoring and interventions and a higher number of surgical births means that more staff are needed than if the same women were having spontaneous, physiological births. Already-depleted staffing levels can mean that entering the maternity services to have a baby becomes more dangerous to all women.

Add in systemic racism, and the danger increases further to Black and Brown women, who research has shown are less likely to be listened to and more likely to be dismissed in systems of care that were built for and by white people.

But if you’re an individual woman and not a policymaker, you don’t need to worry about the other 359 women might want. You just need to make the decision that’s right for you.

So do these data show that induction at 39 weeks beneficial?

I don’t think so, and I’ll explain just a couple of the reasons why.

First, this study used a composite measure to define ‘adverse perinatal outcome’.

Composite measures are tricky things. The composite measure also includes babies who had certain procedures, or who were admitted to the neonatal unit, so some of the babies deemed to have an adverse perinatal outcome will be alive and well. Some may never have been ill in the first place, as I will explain in a moment. Having an adverse perinatal outcome doesn’t necessarily mean that a baby died or even was very unwell.

The use of a composite outcome can sometimes make it appear that there is a significant and genuine difference between two different paths, when this is not the case.

Sometimes babies are given procedures ‘just in case,’ but they didn’t have a problem. In this situation, it can be a good thing that we tested, because it’s sometimes better to err on the side of caution, but having a certain kind test in itself doesn’t mean that a baby was diagnosed with a medical condition. It might in some cases simply mean your baby had a new, inexperienced, or anxious paediatrician. Or perhaps the guidelines insist on offering testing, even when there is no clinical sign of a problem.

There are also situations where some of the women who declined intervention may find that they are told that their babies need testing. This happens not because there is an actual problem, but because staff are fearful of situations that go ‘outside the guidelines’. It also sometimes happens that babies are ‘suspected’ to have a condition that they do not actually have.

A good example of this is where women decline testing and/or antibiotics for group B strep. In some areas, women who decline this may be told that their babies need to be admitted to a neonatal care unit for testing. Even where the baby is fine, that admission will be recorded on their medical records. Then, when researchers come along later who have defined ‘neonatal unit admission’ as one of the criteria in their ‘composite adverse outcome score, they are going to record that baby as having an adverse outcome. Let’s say that the same mum declined induction … hopefully you are starting to see why we need to look more deeply at some of the issues to do with how the research itself is carried out.

So when it says that 360 inductions are needed to avoid one adverse perinatal outcome, that’s not the same thing as saying that 360 inductions will save the life of one baby. An adverse outcome can include a procedure or a diagnosis of something that will not necessarily even affect the baby in later life.

That’s why it’s so important to understand the bigger picture.

What are we counting, and when are we counting from?

My second concern is about what we are counting, and how. This is something I have written about in a few of my books.

For example, the way in which we record stillbirths varies between studies and it’s not always possible to know whether stillbirth happened before or during labour. If it happened before labour, then there is no chance that induction would have prevented it.

Let’s look at the actual numbers in this study:

The risk of any adverse perinatal outcome was 3.28% for the IOL group and 3.64% for the expectant management group (p < 0.001, Table 2).” (Muller et al 2023).

The difference between 3.28% and 3.65% is tiny. Yes, it can be shown to be statistically significant, but this shows exactly why I always recommend looking at the absolute risk and not taking comments about risk reduction or relative risk at face value.

But there’s more.

“The types of adverse perinatal outcomes were also different. Pregnancies in the IOL group had a lower risk of stillbirth (0.01% versus 0.07%, p < 0.001) and other adverse perinatal outcome (3.18% versus 3.53%, p < 0.001) but a higher risk of neonatal death within 28 days (0.10% versus 0.04%, p < 0.001).” (Muller et al 2023).

This is something that I discuss in In Your Own Time, because a few studies have now shown the same thing. Induction can sometimes be shown to reduce the chance of stillbirth in some studies, albeit by a miniscule amount and at a cost, but it can lead to higher rates of neonatal death.

How to find out more

It’s so important to look beyond the headlines, and to not take what people are telling you at face value.

I’m aware that there is much more that I could say about this and similar papers, but I have said a lot of it in my book, In Your own Time: how western medicine controls the start of childbirth and why this needs to stop, so I won’t repeat it here.

And I will mention about that I have an entire hub of induction-related resources on this website.

I hope they help you make the decisions that are right for you.


Want to stay up to date and get regular updates on birth-related research and thinking sent free to your inbox?

Sign up to my newsletter list.


Books to help you become informed and make the decisions that are right for you…


If this blog post or any of my work has helped you or those you care for, please buy me a coffee and help me continue to provide independent birth information online as well as in my books.

Thank you 💜