Might there be risks of risk-based language?

Six years ago, I wrote a blog reflecting on Diane Wiessinger’s seminal essay, “Watch your language.” “There are no benefits of breastfeeding,” I wrote. “There are risks of formula feeding.”

That post remains the most-viewed piece I’ve ever written, with more than 74,000 views as of this writing. I’ve taken the lesson to heart. I’ve published a peer-reviewed study on the increased risk of hypertension among women with curtailed breastfeeding, and I’ve flipped odds ratios in teaching slides and review articles to frame associations as the “risk of not breastfeeding” or the “risk of formula,” rather than the “benefits of breastfeeding.”

Weissinger’s 1996 essay rests on the position that breastfeeding is the physiologic norm, against which all other feeding methods should be compared. Moreover, she notes, mothers who are facing difficulties will be more likely to seek help to avoid a risk than to achieve a benefit:

When we fail to describe the hazards of artificial feeding, we deprive mothers of crucial decision-making information. The mother having difficulty with breastfeeding may not seek help just to achieve a ‘special bonus;’ but she may clamor for help if she knows how much she and her baby stand to lose.

Thus, when we talk about risks of formula, we will motivate mothers to “clamor for help,” and thereby increase breastfeeding rates and improve the health of mothers and babies.

It’s a compelling logical argument. And yet, I’ve been unable to find empirical evidence that it is true. To generate that evidence, we’d need to compare outcomes among mothers and babies counseled that formula increase risk with outcomes among those told that breastfeeding improves health and wellbeing. To my knowledge – and please let me know if there is a peer-reviewed study out there! – such a study has not been done.

In the absence of such evidence, I’ve become increasingly concerned that risk-based language might do more harm than good. In her spectacular keynote address at Breastfeeding and Feminism, Kimberly Seals Allers used a baby bottle with the head of Darth Vader to frame a discussion about how “risks of formula” might be perceived among women of color. For a mother worried that her teenage son might be gunned down for walking down the street wearing a hoodie, “risks of formula” seem pretty trivial. Furthermore, Seals Allers noted, invoking medical experts may ring hollow for communities that remember the legacy of Tuskegee and endure routine undertreatment from medical professionals. Instead, Seals Allers suggests that moms taste infant formula, or read the list of ingredients, and contrast that with mom’s own milk. Similarly, in her address at the First Food Forum, she notes that for her, breastfeeding wasn’t about the health benefits, much less avoiding the risks of formula – it was about rewriting the narrative. (If you haven’t yet seen her talk, stop reading, and watch the YouTube Video here.)

Moreover, there’s mixed evidence that fear changes health behavior, particularly for individuals who lack self-efficacy. In an analysis of six studies that measured threat and self-efficacy, Peters and colleagues considered the impact of threat-based health messages on people with high vs low self-efficacy. Among those with high self-efficacy, threat-based messages worked well – if I learn that sedentary behavior increases my risk of dying, and I can go for a run every morning on a safe, well lit route, and have time to shower and dress before my work day begins, then the threat of premature cardiac death may be effective to get me to work out. But if my home faces a busy roadway with a 55 MPH speed limit, no shoulder and no sidewalks, and I can barely punch in on time at my low-wage job after my kids board the school bus, it’s a very different story.

In a breastfeeding context, messaging that formula feeding increases childhood leukemia risk is unhelpful for the 22% of employed women who return to the workforce by 10 days postpartum, or for the mother who lives 3 hours from the nearest lactation consultant. If the purpose of risk-based language is to motivate mothers to clamor for help, there has to be help within clamoring distance.

Indeed, evidence suggests that when the recommended health behavior is not possible, threat-based messaging might actually INCREASE the undesired behavior, because people shut-out messages that they cannot act upon. The figure shows how Peters and colleagues conceptualize this in their paper.

The left graph shows the effect fear appeals have in theory (the black line represents the effect under high efficacy, the grey line the effect under low efficacy, and the dotted line is the main effect observed when efficacy levels are disregarded); the right plot shows the distribution of threat and efficacy associated to a random sample of behaviour-population combinations. (Peters el al)

This is of particular concern for the effects of risk-based messaging on health equity, because many women of color live in First Food Deserts that lack support and resources for breastfeeding. In this context, it’s plausible risk-based messaging might turn them off from breastfeeding entirely.

The message that formula is risky is further fraught by the fact that a substantial proportion of infants born in the US require supplementation. Delayed onset of lactogenesis is common, affecting 44% of first-time mothers in one study, and 1/3 of these infants lost >10% of their birth weight. This suggests that 15% of infants — about 1 in 7 breastfed babies — will have an indication for supplementation. If we counsel women prenatally that Just One Bottle will permanently alter the infant gut, then we set up 1 in 7 mothers to believe the horse is out of the barn in the first week of life. A mother might even conclude that the damage has been done, so she might was well wean altogether.

In my own clinical work, I see another potential unintended consequence of risk-based language. Many of my patients are coping with excruciating pain, minimal milk production, or postpartum depression. Some have internalized the “risk of formula” messaging such that they are clenching their teeth and dreading every feeding, yet soldiering on because they believe that “every time I feed formula, I am poisoning my baby.” While one might argue that “risk based language is working,” because this mother-baby dyad is still exclusively breastfeeding, it’s not at all clear to me that risk-based language has improved overall health and well-being. Indeed, if we are going to study the effectiveness of risk-based messaging, we can’t just measure breastfeeding duration and intensity – we need mother-centered measures of a woman’s breastfeeding experience. “Although I never fed my baby formula, I hated every second” should not be considered a good outcome.

In reflecting on the potential risks of risk-based language, I was fascinated to read an essay by Robert J. Vallerand, ‘On the psychology of passion.’ Vallerand notes that individuals can experience passion that is harmonious, or that is obsessive. For example, an injured athlete with harmonious passion will sit out her daily run to let an injury heal, whereas an athlete with obsessive passion will keep training through the pain, potentially worsening the injury.

My UNC colleague Barbara Fredrickson has suggested that whether we develop obsessive or harmonious passion relates to the positive or negative emotions we experience with an activity. Positive emotions encourage us to engage with people around us, seek support and build resources, and develop harmonious passions. Negative emotions, on the other hand, foster fear, isolation, and obsessive passion. This body of work has led me to worry that fostering “fear of formula” to motivate breastfeeding may isolate mothers and engender obsessive passion around infant feeding.

This issue of obsessive passion seems to be particularly prominent among middle class, privileged women. As my colleague Cecilia Tomori pointed out to me, the mandate to be “effortlessly perfect” is especially prominent among white middle-class women. I remember agonizing over parenting books filled with acronyms and regimens, desperate to provide a “perfect” environment for my firstborn child. I nodded along with a recent essay, “Motherhood Helped Me Let Go of Perfection.” And yet, I am struck that my life was pretty insulated for me to make it to motherhood without letting go of perfection. As a Native American lactation consultant noted at the First Food Forum, “My clients are not freaking out if they have to supplement a baby with formula.” As we reconsider how to frame the health impact of breastfeeding, we must take care not to center the discussion on how language about “risk” or “benefit” affects well-to-do privileged women.

I’ve also thought about the risks of risk vs. benefit language when talking with health professionals. Shouldn’t we use breastfeeding as the normative standard in that context? Here, again, I’ve not found a study. It would be ideal to compare a health provider’s motivation to engage with patients after hearing a talk on the “risks of formula” vs the “benefits of breastfeeding.” However, my experience is that in every audience, there are parents, and there are friends and relatives of parents. For many, the topic of infant feeding stirs deep emotions. And when we frame the conversation about “risks of formula,” there’s a real risk that we will trigger negative emotions that will cause those in the audience to shut down, such that they will not be able to hear the evidence-based, clinically applicable messages that could enable their patients to achieve their own infant feeding goals.

Finally, it’s critical that however we seek to motivate families to breastfeed, we pair that motivation with a much bigger dose of practical, actionable ability to initiate and sustain breastfeeding. In the book “Influencer: How to Change Anything,” Joseph Grenny and co-authors note that to change a behavior, an individual has to answer “yes” to two questions: “Is it worth it?” and “Can I do it?” Far, far too often, we plead, beg, bludgeon and entice people to change a health behavior, with minimal consideration for whether that behavior is actually possible. The futility of that strategy is reflected in steadily rising breastfeeding initiation rates in the US, followed by a precipitous drop in the first weeks after birth. The vast majority of families want to breastfeed – the problem is the barriers that prevent them from doing so.

My sense is that when we remove the “Booby Traps” that undermine so many women, and when we ensure that every mother and baby had access to appropriately trained lactation support providers, long term breastfeeding rates in the US will soar, regardless of whether we frame health outcomes as “risks of formula” or “benefits of breastfeeding.”

And until we have evidence that risk-based language improves the wellbeing of mothers and children, we might do well to reconsider our rhetoric. It’s past time to conduct the research that’s needed to sort the real risks and benefits of risk-based language.

Alison Stuebe, MD, MSc, is a maternal-fetal medicine physician and breastfeeding researcher. She is an associate professor of Obstetrics and Gynecology at the University of North Carolina School of Medicine and Distinguished Scholar of Infant and Young Child Feeding at the Gillings School of Global Public Health. You can follow her on Twitter at @astuebe.

Posts on this blog reflect the opinions of individual ABM members, not the organization as a whole.

Addendum, posted 4-15-16, 8:15 am

This post has inspired spirited discussions on a number of forums, and one of the refrains I’ve heard from many is that health care providers would never talk about the “benefits of not smoking” – they would talk about the risks. One commenter on a Facebook group stated, “Humans by nature ARE risk-averse, it’s psych 101. We act to avoid risk, not accrue benefit.”  Why, many have asked, would we withhold from families that there are risks of formula, compared with the biological norm of breastfeeding?

With that question in mind, I reviewed the evidence-based recommendations for smoking cessation among pregnant women, and I was struck by how providers are advised to encourage women to quit smoking:

Advice to quit should be clear, strong, and personalized with unequivocal messages about the benefits of quitting for both the patient and her baby. An effective way to start the discussion about quitting is to say, “My best advice for you and your baby is for you to quit smoking.”
Additional advice can then be tailored to the patient’s situation and their responses to the multiple choice “Ask” Question, using positive language and focusing on the positive benefits of quitting. Although clinicians are keenly aware of the danger smoking poses to infants and the longterm health risks for mothers, it is common for patients to minimize risks.
Focusing on bad outcomes such as low birth weight or delivery complications may be ineffective for patients who believe they are not at risk, especially if they or people they know have had uncomplicated, healthy pregnancies while smoking. Describing the good things the patient can do for herself and her baby by quitting smoking appeals to her desire to be a good mother. Table 3 includes examples of benefits of quitting that clinicians can use when advising patients

When you stop smoking…

• your baby will get more oxygen, even after just one day of not smoking
• your baby is less likely to have bronchitis and asthma
• there is less risk that your baby will be born too early
• there is a better chance that your baby will come home from the hospital with you
• you will be less likely to develop heart disease, stroke, lung cancer, chronic lung disease, and other smoke related diseases
• you will be more likely to live to know your grandchildren
• you will have more energy and breathe more easily
• you will have more money that you can spend on other things
• your clothes, hair, and home will smell better
• your food will taste better
• you will feel good about what you have done for yourself and your baby

(Italics mine; Reference: Page 8, Smoking Cessation During Pregnancy: A Clinician’s Guide to Helping Pregnant Women Quit Smoking )

These recommendations are based on studies that have tested the effectiveness of risk vs. benefit language for encouraging women to stop smoking — a behavior that is universally understood to be unhealthy — and the recommendation is NOT to focus on the risks, but to emphasize the benefits, to motivate women to change behavior.  While we may believe that humans act to avoid risk, the evidence from smoking cessation suggests that we change health behaviors to accrue benefit.

I am not suggesting that we should banish risk-based language from every conversation about infant feeding. What I am suggesting is that we’ve embraced risk-based language based on a rhetorical position that framing formula as risky will inevitably make breastfeeding the cultural norm.  This argument assumes that individuals will react to messages about the risk of formula by enacting behaviors and policies to enable all families to breastfeed — and we simply do not know if that is true. Those of us who are already actively engaged in breastfeeding work may be motivated by risk-based framing, but our intended audience is not breastfeeding activists.

The fundamental question we have to ask is “What Do We Really Want?” Do we want to broadcast the message that breastfeeding is normal, and formula is risky, even if that message might be ineffective for a substantial proportion of the population? Or might we want to find an approach meets each family where they are and engages them on their terms?