Lead
Breast milk or formula? If you have a baby under six months, there's a good chance you've typed some version of that question into a search bar more than once.
What comes back is rarely calm. "Formula lowers IQ." "Breastfeeding is exhausting and unnecessary." "Mixed feeding is the worst of both worlds." Each claim arrives with a plausible-sounding number or a personal story attached.
The problem isn't that these statements are made — it's that most of them strip evidence from its context. "Domains where breastfeeding matters" and "domains where the effect is barely detectable" are different conversations. And associations observed in cohort studies: observational research that follows a group of people over time to see who develops an outcome, without randomly assigning any treatment are not the same as effects established in randomized trials.
This article draws on the 2016 Lancet Breastfeeding Series and the PROBIT trial, among other peer-reviewed sources, to map the axes you actually need to compare when making this decision. No conclusion is offered. That's not the kind of question that has one.
Short-term and long-term evidence are not equally strong
Start with the broad picture. The evidence base for breastfeeding benefits follows a clear structure: it is considerably stronger for short-term outcomes (infections, infant mortality) and weakens as you move toward long-term outcomes (IQ, obesity, metabolic disease).
The 2016 Lancet Breastfeeding Series (Victora et al.) synthesized data from low-, middle-, and high-income countries in a large meta-analysis, finding that breastfeeding significantly reduces rates of infant gastrointestinal infection, respiratory illness, and mortality [1]. Specifically, longer breastfeeding duration was associated with a 26% reduction in childhood overweight and obesity risk (95% CI 22–30%), and the authors estimated that improved breastfeeding practices could prevent approximately 820,000 deaths of children under five annually worldwide [1].
The persuasiveness of those numbers, however, depends heavily on context. The protective effect against infant mortality and diarrheal illness is documented most strongly in low- and middle-income countries with poorer sanitation. Extrapolating the same effect size directly to a high-income country population — including countries like Japan with clean water and formula quality regulation — is not straightforward. The Lancet series itself is careful to draw this distinction [1].
Moving to long-term outcomes, the evidence weakens further. A systematic review and meta-analysis by Horta, Loret de Mola, and Victora (2015) pooled 17 studies and reported that breastfed children scored an average of 3.44 IQ points higher [2]. However, in the subset of analyses that adjusted for maternal IQ, the difference shrank to 2.62 points [2]. The confounding question — do higher-IQ mothers breastfeed more, and also pass on higher IQ genetically? — remains only partially resolved.
What PROBIT showed: the gap between intervention effect and observed association
No study is cited more often in the breast-milk–versus-formula debate than the PROBIT trial by Kramer and colleagues [3]. PROBIT cluster-randomized 31 maternity hospitals in Belarus to either a Baby-Friendly Hospital Initiative breastfeeding promotion intervention or standard care, and followed 17,046 children to age 6.5 years [3].
Children in the intervention group scored 7.5 points higher on the verbal IQ scale and 5.9 points higher on full-scale IQ of the Wechsler intelligence test, with teachers also rating them higher on academic performance [3]. Because this is a randomized trial, it sidesteps at least some of the "breastfeeding mothers were already more educated and affluent" confounding that plagues observational work.
Two features of PROBIT are frequently overlooked, however. First, what was randomized was not "breastfeeding" but rather a complex hospital-based intervention — the Baby-Friendly Hospital Initiative — that includes lactation support, maternal education, and increased skin-to-skin contact. While the intervention did raise breastfeeding rates, it is not possible to isolate the effect of breast milk per se from the broader package of support and education. Second, the 16-year follow-up of PROBIT adolescents found that the cognitive advantages seen at 6.5 years had partially attenuated by adolescence [4].
In short: claiming that "breastfeeding substantially raises IQ" overreads what PROBIT found, but so does claiming "breastfeeding has no effect on cognition." At the population level, a small-to-moderate difference is observed. It is difficult to attribute that difference cleanly to breast milk alone, and it diminishes over time. That is approximately where the evidence stands today.
The parent's health — the axis that often goes missing
This is where the decision becomes most personal. Discussions of breast milk versus formula that focus exclusively on infant outcomes leave out an essential second axis: the physical and psychological burden on the nursing parent.
A comprehensive review by Stuebe documents that breastfeeding is associated with long-term maternal health benefits, including reduced risks of breast cancer, ovarian cancer, and type 2 diabetes [5]. At the same time, the clinical reality of lactation — sleep deprivation, mastitis: a painful infection of breast tissue that can occur during breastfeeding, causing redness, warmth, and fever, engorgement, and the self-blame that often accompanies difficulty with milk supply — is well recognized as a contributor to postpartum depression [5].
The decision about feeding method therefore involves at minimum four dimensions:
- Infant short-term outcomes (infection risk reduction) — relatively strong evidence, though absolute risk differences are smaller in high-income countries
- Infant long-term outcomes (IQ, obesity) — weak-to-moderate evidence, with residual confounding
- Parent's long-term health — moderate evidence for benefits
- Parent's physical and psychological burden — highly individual; not negligible
Weighing these axes against each other is not something evidence can do for you. One parent may find the fourth axis most decisive. Another family may weight the first most heavily. Which axis matters most is a values judgment, and only the person living with the decision can make it.
The "absolute breastfeeding" and "formula is fine, stop the guilt-tripping" camps that circulate on social media both tend to hold up one of these four axes while quietly setting down the other three. That's understandable as advocacy. It is not a framework for your own decision-making.
The Japanese numbers and the reality of mixed feeding
A useful reference point: according to the Ministry of Health, Labour and Welfare's 2015 Infant and Young Child Feeding Survey, at one month of age 51.3% of infants in Japan were exclusively breastfed, 45.2% were receiving mixed feeding (breast milk and formula), and 3.6% were exclusively formula-fed [6]. At three months, the figures were 54.7%, 35.1%, and 10.2% respectively [6].
What stands out is that mixed feeding — often portrayed as a compromise or inconsistency — accounts for roughly a third to half of all infants at any point in the first few months. The breast-or-formula framing that dominates online discussion does not reflect how most Japanese families actually feed. The 2016 Lancet series frames its findings around duration and degree of breastfeeding on a continuum, not as a binary [1]. The data suggest that treating it as one is a distortion to begin with.
Three practical steps
Three things worth considering — offered as independent options, not instructions.
First, create some distance from the extreme voices on social media. Both unconditional breastfeeding advocacy and dismissive formula-promotion tend to blend personal justification with selective evidence. When you trace those claims back to the peer-reviewed literature, the effect sizes are more modest and the uncertainty is larger.
Second, articulate, in your own words, how you weigh the four axes for your family. "Getting enough sleep matters most for my mental health." "I want to preserve time for my older child." "In our home environment, infection risk differences feel small." Your priorities are yours to set. A record-keeping app like Memori can hold feeding logs alongside notes on your own physical state and mood — so that when you look back, you can see how your priorities shifted over time, not just what you fed.
Third, if you are struggling, contact a lactation consultant or midwife early. Spending more than a few weeks feeling that breastfeeding "isn't working" — without support — has consequences for milk supply, and for mental health. Earlier is better. There is no such thing as consulting too soon.
Summary
"Breast milk or formula?" is a multi-dimensional question that bundles evidence strength, applicable range, parental burden, and family values — not a question that yields a one-sentence answer.
Long-term effects on IQ and obesity risk are suggested by observational data but tend to diminish in randomized trials [2,3,4]. Short-term infection protection is comparatively robust but the absolute risk difference is smaller in high-income country populations [1]. Any recommendation that ignores the mental and physical toll on the nursing parent is reading only half the evidence base [5].
The weighting is yours to do. That's precisely why you don't need to internalize the certainties of people who aren't living your situation.
References
- Victora CG, Bahl R, Barros AJD, et al. Breastfeeding in the 21st century: epidemiology, mechanisms, and lifelong effect. Lancet. 2016;387(10017):475–490. doi:10.1016/S0140-6736(15)01024-7. PMID: 26869575.
- Horta BL, Loret de Mola C, Victora CG. Breastfeeding and intelligence: a systematic review and meta-analysis. Acta Paediatr. 2015;104(467):14–19. doi:10.1111/apa.13139. PMID: 26211556.
- Kramer MS, Aboud F, Mironova E, et al; Promotion of Breastfeeding Intervention Trial (PROBIT) Study Group. Breastfeeding and child cognitive development: new evidence from a large randomized trial. Arch Gen Psychiatry. 2008;65(5):578–584. doi:10.1001/archpsyc.65.5.578. PMID: 18458209.
- Yang S, Martin RM, Oken E, et al. Breastfeeding during infancy and neurocognitive function in adolescence: 16-year follow-up of the PROBIT cluster-randomized trial. PLoS Med. 2018;15(4):e1002554. doi:10.1371/journal.pmed.1002554. PMID: 29677187.
- Stuebe A. The risks of not breastfeeding for mothers and infants. Rev Obstet Gynecol. 2009;2(4):222–231. PMID: 20111658.
- Ministry of Health, Labour and Welfare, Japan. Summary Report of the 2015 Infant and Young Child Feeding Survey. 2016. https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/0000134208.html
- Rollins NC, Bhandari N, Hajeebhoy N, et al; Lancet Breastfeeding Series Group. Why invest, and what it will take to improve breastfeeding practices? Lancet. 2016;387(10017):491–504. doi:10.1016/S0140-6736(15)01044-2. PMID: 26869576.