In 1996, Diane Wiessinger published “Watch Your Language”: an editorial for the Journal of Human Lactation, in which she argued that:

Artificial feeding, which is neither the same nor superior, is therefore deficient, incomplete, and inferior”.

This rallying cry to breastfeeding advocacy essentially proposed that the conventional approach to talking about the benefits of breastfeeding was an “inversion” of reality. She proposed that because breastfeeding is the biological norm, all statistical representations presented to parents to help with their decision making should use breastfeeding as the referent. This would therefore mean that statistical representations state that babies fed formula are “more” at risk of a particular harmful health outcome (eg, diarrhea), not that breastfed babies are “less” at risk. Wiessinger explains that this “warning” language would be sufficiently terrifying to parents to compel them to seek whatever help they could to ensure that breastfeeding was successful.

In the twenty-two years since this article has been published, it has had great influence (dozens of citations on Google Scholar), as well as being implemented into information put forward directly to families, such as:

  • Mass media articles: “Rather than seeing breast as best… breast is normal. “It’s what your body is normally designed to do,” she says. So, breastfeeding doesn’t reduce the chances of a baby becoming obese; it’s formula feeding that increases the chances of childhood obesity…. we should be talking about the risk of formula feeding and not the benefits of breastfeeding”
  • Statements by policy makers: A widespread misconception by almost everyone that formula milk can replace breast milk without any harm”
  • Popular resources used by parents: Keep in mind that formula (like a medication) comes with its own set of risks… The use of formula can shorten the breastfeeding relationship, and comes with increased health risks for baby and mother”
  • Guidance for healthcare commissioners (p3): NOT breastfeeding is linked to an increased risk of…”

And, of course, tweets:

I wanted to understand whether this type of representation is useful in public health terms. Despite counter-arguments, such as this one from a respected US obstetrician-gynecologist and breastfeeding researcher and advocate, it is still common to see “formula->risk” (or “not breastfeeding->risk”) framing. Indeed, while training in voluntary breastfeeding support I was once told that it was not correct to say breastfeeding had benefits. Rather, formula had risks. (I did try to argue against this, and I think my trainer acknowledged my concerns, but since then I have seen widespread examples elsewhere of the same framing).

As I explain below, there is evidence from a variety of sources that risk-based language could be counterproductive and actually end up undermining the overall public health goal – that is, to benefit the health of mothers and babies. Here are five reasons why I think we should step away from risk-based language in infant feeding:

1: Statistics doesn’t care which referent you choose

Wiessinger’s original essay argues that “Health comparisons use a biological, not cultural, norm, whether the deviation is harmful or helpful”. However, having checked all the medical statistics, epidemiology, and evidence-based medicine textbooks that I can lay my hands on, I could find no particular opinion on this proposal. Even the incomparable Trish Greenhalgh (How to Read a Paper) and Hayes, Sackett, Guyatt and Tugwell seem to say nothing on the matter. Finally after much searching I did find some comment on the issue: a statement in Oakes and Kaufman, “Methods in Social Epidemiology”, which says (in relevance to the choice of referent group, in general for epidemiological studies):

“There is no “correct” reference group, but several choices are possible, and it is perhaps most important to make the rationale for the choice of referent clear”.

Therefore a utility argument should apply: the best choice of referent should surely be the one that in practice, results in the best health outcomes for the target population – which may be the language that best encourages parents to engage openly with, and get help from their healthcare professionals on issues of infant feeding.

2. Humans are deeply cultural: there is no “biological norm”

Moche woman breastfeeding. Pattych,  en.wikipedia, CC BY-SA 3.0

Discussions about the “risks” of infant formula often use as their starting point an understanding that breastfeeding would be the “biological norm” for humans. In her original essay, Wiessinger doesn’t define what “biological norm” means, but we can assume that she is referring to whatever processes or behaviours existed for ancient (but biologically modern) humans before we had all the trappings of contemporary society. The problem is, there was no such thing. Human society is hugely diverse and humans have been inventing and transmitting cultural and technological innovations over millennia as they have adapted to the challenges of their surroundings. A fantastic blog by a US professor of anthropology explains that, for humans, there is no bright line between “biology” and “culture”: we are our culture, and anthropology seeks to understand and explain everything from Cave Mama to Lady Gaga:

“The comparative work of Cultural Anthropology leads us to insist that no one form of human life is “natural.” There are no bedrock or natural roles for sexes and genders; for parenting; for kinship; for economic systems; for political organization. The social inequalities we observe are not inevitable. We can live differently, whether that means providing healthcare as a human right or curtailing gun violence”.

In addition to that, cultural (or technological) innovations allowing us to depart from what our biology would otherwise support have been in existence even before the earliest Homo sapiens  – inventions though to have been developed by pre-sapiens hominids include the control of fire; tool use; development of seaworthy craft; clothing; and even potentially sewing needles.

Another way to think about this is what constitutes “normal” human behaviour (and breastfeeding is a behaviour involving two people, with the baby’s reflexes and innate instincts, working together with the mother’s learnt skills). Wikipedia explains that “normality” for human behaviour is hugely varied and the definition of what is “normal” is linked with what someone’s social group would expect, rather than any biological absolute, noting:

“Normal is also used to describe individual behaviour that conforms to the most common behaviour in society…. definitions of normality vary by person, time, place, and situation – it changes along with changing societal standards and norms”.

Consequently, trying to think about human behaviours in terms of “biological norms” leads to a reductio ad absurdum. Yes, breastfeeding is a completely normal behaviour, and should be supported in all human societies. But that is not the same as saying it is THE norm, i.e. no other mode of feeding can be supported by society, or that not doing it indicates that some biological process is not working properly.

Animal shaped baby drinking vessel, late Bronze Age. Bullenwächter – Historisches Museum Regensburg. CC BY-SA 3.0,

From a societal and cultural perspective therefore, there’s nothing particularly unique, technologically, about infant formula in principle that distinguishes it from any other human invention that we take for granted (from the mundane to the miraculous): toilets, clothing, built shelters, vaccines, writing, or the internet. And of course, histories of infant feeding make it clear that since time immemorial, mothers have supplemented their own milk with anything going: unmodified animal milks, prelacteal feeds (water, honey etc), wet nursing and milk sharing; these options are not considered safe or acceptable and are no substitute for the mother’s own milk, infant formula, or screened, banked donor milk. But what this shows is that formula is just a safe, nutritionally adequate variant on a cultural continuum that has been in existence as long as humanity. (nb, when I say safe and nutritionally adequate here, I am referring to the UK and similar settings).


3. It doesn’t reflect what we know about the causal pathways underlying health differences

If information targeted to parents uses the “formula=risk” approach to framing, this may result in parents believing that formula directly causes harmful outcomes, i.e. disease in babies. Yet, this would be an inappropriate conclusion from the actual science underlying this field, for the most clearly understood health outcomes that have been studied. For example, some of the strongest evidence for positive effects of breastfeeding on infant health in developed countries are for infant infections (eg, respiratory tract infections, otitis media, diarrhea etc). But infant formula does not cause a respiratory tract infection. Rather, the mechanism of effect is that breastfeeding is protective, i.e. it prevents infections that would otherwise occur in babies.

These mechanisms are fairly well understood, and stem from the transfer via mother’s milk of IgA and IgG antibodies, bioactive factors such as proteins (eg, lactoferrin), cellular components (eg, macrophages, lymphocytes etc), and many other immune components (Lawrence and Lawrence, pp148-176). Collectively these constituents in human milk and colostrum “mop up” and inactivate bacterial and viral infective agents that might otherwise cause disease. However, respiratory tract infections, otitis media, and diarrhea in infants in developed countries can be caused by a huge range of bacteria and viruses and many of these are transferred (as any mother will know, the writer included!) from person to person – typically from one snotty nosed child to another snotty nosed child, via droplet infection or other routes.

A causal pathway of “formula->harm” is only really plausible in developed countries, for diarrheal disease resulting from contaminated or poorly made up formula, as the Lactalis scandal has shown; this highlights why strong regulation and food safety protection are important, and why it’s crucial to make sure that formula is properly made up according to guidance. However, in general it would not be correct to say that formula actually causes the health differences seen in epidemiological studies comparing formula-fed versus breastfed babies; rather, breastfeeding helps to prevent the common infant infections for which risk differences appear. In very poor countries, where safe water supplies often do not exist and many families may have no way to make up formula properly, I do believe it is OK to say formula is risky to infant health, because this would be a reasonable way of stating what the science tells us for those settings.

4. Risk-based approaches to communication may drive behavior in the opposite direction to that intended

As I explain above, I generally think that the issue of which “risk framing” to choose boils down to a utility argument: whatever approach results in the best outcomes in the real world, is the one that should be adopted. There’s a good example of this in Gerd Gigerenzer’s book, “Risk Savvy” (p5-6), when he talks about the Pill Scare:

“the UK Committee on Safety of Medicines issued a warning that third-generation oral contraceptive pills increased the risk of thrombosis twofold – that is, by 100%… Alarm bells rang… Distressed women stopped taking the pill, which caused unwanted pregnancies and abortions. The absolute risk increase was only one in seven thousand. This single scare led to an estimated thirteen thousand additional abortions in the following year in England and Wales… for every abortion there was also one extra birth… Ironically, pregnancies and abortions are associated with a risk of thrombosis that exceeds that of the third-generation pill”

What Gigerenzer is saying here is that our efforts to warn people about risk may actually result in MORE of the risk we are trying to avoid than would have otherwise happened. Which is not to say that we should not give people information, rather we should give them information presented in a way they can interpret and use.

For breastfeeding, the public health goal here surely is that we should present information framed in a way that motivates women who already want to breastfeed to seek the support they need to continue, while not alienating women who don’t want to breastfeed in the first place.

Helpfully, data is available on “what works” in the real world. Two studies have been conducted; one involving expectant mothers, and the other undergraduate students, both male and female. In both studies, the participants were randomized to differently framed (positive or negative) but otherwise identical, presentations of the same information about how infant health outcomes may be affected by breastfeeding or formula feeding. The two studies had completely consistent results: negatively framed presentations seemed to alienate participants from breastfeeding promotion strategies. In the “expectant mothers” study, intent to breastfeed was not altered by use of “negative” risk framing but these respondents had less favourable assessments of the information provided.

The researchers say:

“We call for breastfeeding advocacy and health promotion strategies that respect the autonomy and intelligence of mothers. A woman-centered approach trusts mothers to do the best they can for their children in the social and economic contexts in which they parent. Mothers’ rejection of risk language breastfeeding advocacy demonstrates the ineffectiveness of advocacy methods that do not include women as equal partners with health advocates”.

Qualitative research from the UK reinforces this idea about moving breastfeeding advocacy away from idealism and towards realism – supporting the individual with their own choices in their own unique situation, rather than viewing anything different to guidelines as “inferior” or a “risk”. In their study, Hoddinott and colleagues describe observing dichotomies of ‘good’ or ‘bad’, ‘right’ or ‘wrong’, ‘breast or bottle’ prevail[ing] in accounts of interactions [between parents and] health professionals”.

5. It’s all connected: policy to grassroots, via social media

One of the most worrying things about the “risk based” language that I discuss here is that calls to use it are often not just isolated, rare examples, but sometimes are set out by those with greater involvement and influence, e.g. with regard to policy development. My concern therefore, is that mothers might ultimately be alienated, not just by specific examples of this “risk language”, but by the wider community that is involved in breastfeeding promotion and support – potentially even healthcare professionals involved in her baby’s care. This would be a massive shame, because of course those professional and volunteer communities are incredibly important in providing support for infant health directly to families.

A study carried out in a deprived UK area where specialist breastfeeding support was available, but not extensively accessed, looked at why many women did not use the support services available. The researchers explained:

Women sensed pressure… to breastfeed and moral judgement around their feeding decisions. It was felt that women were made to feel ‘guilty and bad’ if they chose not to breastfeed and felt like a ‘failure’ if breastfeeding difficulties arose’.

We cannot say that these concerns, and reasons some mums have for not seeking help are directly linked to “risk based language”, but the findings do highlight links between mothers’ perceptions of moral judgement from those who are hoping to help them, and their decisions to seek help or not.

In a nutshell, mothers who hear the phrasing “formula has risks” may see that as pejorative, whether they are still breastfeeding, previously did and stopped, or never breastfed at all. Conceivably this could deter mothers who have already started with some formula, from seeking help from breastfeeding support groups. Why would a mother go to a group if she thinks that the group leader is going to criticize her parenting?

A recent research study examining the emotional states experienced by 890 mothers feeding their babies formula backs up this concern: the researchers find that 67% of mothers reported guilt, 68% stigma, and 76% said they needed to defend their decision to feed their babies formula. The researchers find that both internal and external motivations were involved, and drivers could include other mothers as well as healthcare professionals:

These feelings may occur as a result of not conforming to health professionals’ recommendations or stem from a perception that health professionals judge formula to be an inferior option”.

These findings completely echo what I experienced at breastfeeding support groups. It was so common for mothers to be desperately sad, breaking down in tears, because they needed to give their babies formula – sometimes just one bottle. Often they had done everything they could to get breastfeeding to work. It cannot possibly be helpful for mothers struggling in this situation to hear about the risks of formula.

Finally, in terms of socioeconomic equity, it has been said that “breastfeeding is a natural safety net against the worst effects of poverty”, but less advantaged women often live in communities where babies are more likely to be formula fed and therefore may be even more alienated by the inclusion of “risk” messaging within breastfeeding promotion strategies. Paradoxically therefore, saying “formula has risks” is potentially a regressive policy – working to deepen the socioeconomic patterning of suboptimal infant health rather than to reduce it.

Moving forwards

Thankfully, not everyone uses this language (although it is more common than you’d think), and some active within the support and advocacy community seem to be working towards change. In a new study, UK researchers plan to evaluate an intervention aiming:

to establish a strong supportive helper–woman relationship with continuity of care from pregnancy until after birth, respect a woman’s choices, and be non-judgemental and offer discussion of breast feeding and formula feeding issues”.

This initiative is important because it avoids dichotomising infant feeding into just two modes (breastfeeding versus formula feeding), recognizing that infant feeding decisions exist on a continuum and because it also acknowledges that formula-feeding parents, as well as breastfeeding mothers, are under served. A recent UK report revealed that 67% of midwives did not have enough time to help formula-feeding parents with the information they needed, while only 29% felt they did not have enough time to help new mothers with breastfeeding.

Importantly, infant feeding is not the only field that is finding itself forced to confront the potential harms of shame-based risk messaging. UK GP Margaret McCartney writes in the BMJ about the recent Cancer Research UK “OB_S__Y” campaign that public health initiatives do not “get to choose the effects of [their] campaigns”. What she’s saying here, is that regardless of what statistical “correctness” underlies the message, if people aren’t impressed, then it’s a dud.

Moving forwards, it’s important to acknowledge the complexity and variability in infant feeding behaviours – aware of the full range of goals that mothers might make, the varying strength of their intentions, and the way in which things can, and need to, change when Real Life wades in and nothing goes to plan. A move away from idealistic, “risk-centred” language, and towards a more individual-focussed mode of discussion would hopefully benefit everyone, as well as being an acknowledgement that public health (the primary argument for breastfeeding promotion and support) is at root a fundamentally pragmatic discipline that aims to help real people in the real world.


Many thanks to Dr Anna Pease and Dr Fiona Woollard, who have been very generous with their ideas and time, and gave suggestions on an earlier version of this blog. Errors, omissions or failures of judgement are all mine.

Note: I would also like to point readers to an earlier blog by “Good Enough Mummy” written some years ago, which I had previously read but only found again during the final stages of rewriting this blog. GEM discusses many of the same ideas, only better.