Browse all notes

Chapter 4 / Part Two / The First Year

Breast Is Best?

What the strongest evidence does and does not show about breastfeeding.

On this page

Core idea: Breastfeeding has some real short-term benefits, but many dramatic long-term claims come from comparing families who differ in many other ways. Feeding decisions should weigh credible benefits against the costs to the parent and family.

One-Minute Recap#

  • A simple comparison of breastfed and formula-fed children does not isolate the effect of breastfeeding.
  • In the United States, breastfeeding is correlated with parental education, income, leave, health behavior, and support. These factors also predict child outcomes.
  • Randomized encouragement studies and sibling comparisons reduce this confounding and find fewer dramatic benefits.
  • The best-supported infant benefits are modest and short term, particularly fewer gastrointestinal infections and possibly fewer rashes.
  • Strong evidence for large effects on IQ, obesity, allergies, or long-term health is lacking.
  • Breastfeeding is associated with a meaningful reduction in the mother's later breast-cancer risk.
  • Enjoyment, convenience, pain, pumping, work, sleep, and mental health are legitimate parts of the decision.

Guidance at a Glance#

Label Takeaway
OPTION - Moderate Breastfeed if the short-term benefits and personal experience make it worthwhile for your family.
OPTION - High Formula feeding is a valid alternative where formula and clean water are reliably available. The book finds no compelling evidence of major long-term harm.
RECOMMENDED - Moderate Support parents who want to breastfeed; practical support can determine whether it is feasible.
RECOMMENDED - Moderate Include the mother's potential breast-cancer benefit in the decision, while recognizing that individual baseline risk varies.
SKIP - High Skip guilt and claims that feeding method determines intelligence, attachment, or future success.
SKIP - High Do not treat an unadjusted observational association as proof that breastfeeding caused an outcome.

Evidence Cheat Sheet#

Claim Book's assessment Why
Less infant diarrhea Likely real, modest benefit Supported by stronger study designs
Less eczema/rash Possible short-term benefit Some trial evidence, but not a universal effect
Higher IQ or school performance Not convincingly causal Effects shrink in randomized encouragement and sibling comparisons
Lower obesity Not convincingly causal Family differences explain much of the association
Lower SIDS risk Uncertain Case-control evidence is vulnerable to confounding
Lower maternal breast-cancer risk Likely real More consistent dose-response evidence across studies

Numbers Worth Remembering#

  • In the large randomized breastfeeding-promotion trial discussed, gastrointestinal illness was roughly 13% versus 9% over the measured period.
  • The maternal breast-cancer reduction cited is roughly 20% to 30%, although individual risk and duration matter.
  • These are population averages, not predictions for a particular parent or child.

Practical Decision Playbook#

  1. Identify the outcomes that actually have credible evidence, not the full list of popular claims.
  2. List the real costs: pain, time, pumping, sleep, medication questions, work constraints, and mental health.
  3. List the personal benefits: closeness, enjoyment, convenience, cost savings, or cultural value.
  4. Decide what level of breastfeeding, combination feeding, or formula feeding is sustainable.
  5. Revisit the decision when circumstances change. Feeding is not a moral commitment or a test of parenting.

What Is Still Uncertain#

  • The precise effect on SIDS, asthma, allergies, and several rare outcomes.
  • Whether very small long-term effects remain after all family differences are addressed.
  • An evidence-defined optimal duration that applies to every family.

My Family Notes#

  • Benefits that matter most to us:
  • Costs or constraints that matter most:
  • Our preferred feeding plan:
  • Our acceptable fallback plan:
  • What would make us revisit it: