Effective Compared With What? Context, Equity, and Delivery

The subject here is not whether breastfeeding support works in an abstract setting. It is what an added service contributes for a particular population, above the care already available, under the social and institutional conditions that make continuation possible or difficult. That perspective helps distinguish a transportability problem—an effect that may not carry cleanly to another setting—from an efficacy problem in the intervention itself.

The baseline is part of the result

The phrase “usual care” can sound like a neutral control condition, but it is actually a local package of services, expectations, and constraints. In the reviewed trials, it could include prenatal education, printed information, hands-on help in the hospital, access to a lactation professional, a telephone line, peer counseling, or postpartum home visits. The exact mix was often poorly described and varied across countries, facilities, and periods. An added program was therefore compared with the care available in that place, not with an identical global standard.

This changes how the pooled effect should be read. If routine care includes little follow-up after discharge, additional contact may have more room to prevent an early problem from becoming a reason to stop. If routine care already offers skilled assistance and a reliable referral route, a new service may produce a smaller difference because both groups receive meaningful help. The smaller difference does not establish that the added service has no value. It may show that the service is filling a gap that the control setting has already partly addressed.

The baseline rate also carries information about the setting. A high initiation rate can leave little room for an intervention to improve initiation, while a lower continuation rate may indicate that families are losing support at a later transition. The same relative risk can produce different absolute gains at different baselines. A program serving a population with lower usual-care continuation may have more potential benefit, but it may also face more severe barriers. Baseline prevalence is therefore not just an input to a calculation; it is a clue about what the intervention is being asked to change.

The population enrolled in a trial is another part of the comparison. The USPSTF evidence base focused on mothers of full- or near-term infants and excluded groups whose medical needs would substantially change feeding care, such as many preterm or medically complex infants. Nearly 44% of the included studies required participants to intend to breastfeed, and most participants in the other studies also intended to do so. That design can be appropriate for testing continuation support among people who have chosen or are considering breastfeeding, but it gives less information about a population in which intentions, medical circumstances, or available options are very different.

The U.S. evidence also has a particular social profile. Thirty-three of the 90 trials were conducted in the United States, and 24 of those mostly enrolled women of color. Many recruited low-income, WIC-eligible, or Medicaid populations, especially Black and Hispanic or Latina women. This representation is valuable because the interventions were tested among groups that have historically experienced lower breastfeeding initiation or continuation and may face barriers to support. It should not be turned into a claim that the pooled effect belongs to every Black, Hispanic, Latina, or low-income family. Within-group circumstances vary, and the study samples themselves were not identical.

Outside the United States, demographic and socioeconomic characteristics were reported less consistently.