Fathers’ Breastfeeding Confidence Falls Into Three Distinct Groups, Shanghai Study Finds

A cross-sectional study of 306 new fathers in Shanghai has found that paternal confidence in supporting breastfeeding is not a single continuous trait but clusters into three distinct subgroups, according to research conducted at a tertiary obstetrics and gynecology hospital affiliated with Shanghai First Maternity and Infant Hospital, Tongji University School of Medicine. The study, published in Frontiers in Nutrition in the journal's nutritional epidemiology section, applied latent profile analysis (LPA), a statistical technique that identifies unobserved subgroups within a population based on patterns of responses. The researchers argue that previous work has largely described paternal breastfeeding support self-efficacy at the group-average level, potentially masking meaningful differences between fathers that could matter for targeted support.

Participants were fathers of newborns recruited by convenience sampling between September 2024 and February 2025. Their mean age was 31.69 years, with a standard deviation of 3.93 years. Each father completed a general information questionnaire, a breastfeeding knowledge questionnaire, and the Father Support Breastfeeding Self-Efficacy Scale-Short Form (FBSES-SF), a validated instrument measuring a father's confidence in his ability to support his partner's breastfeeding.

Across the sample, the mean FBSES-SF total score was 53.03 with a standard deviation of 10.00. Latent profile models were estimated in Mplus 8.3, and the best-fitting solution identified three profiles: a low self-efficacy group containing 57 fathers (18.6 percent of the sample), a moderate group containing 198 fathers (64.7 percent), and a high self-efficacy group containing 51 fathers (16.7 percent). The researchers describe the three profiles as largely parallel in shape and best interpreted as gradations of severity rather than qualitatively different patterns.

To understand what distinguishes membership in these groups, the team used multivariable multinomial logistic regression with the three-step (R3STEP) procedure. This approach is technically notable because it accounts for classification error inherent in latent class models, meaning that the uncertainty in assigning each father to a profile is incorporated into the estimates rather than treating profile membership as if it were directly observed. The analysis yielded several associations with profile membership.

In other words, each additional point of breastfeeding knowledge was associated with slightly reduced odds of falling into the low-confidence group. Attitudes toward breastfeeding also mattered. Fathers who perceived breastfeeding as not important had markedly higher odds of membership in the low rather than the moderate self-efficacy group, with an odds ratio of 9.44 (95 percent confidence interval 1.65 to 53.91).

Conversely, fathers who rated breastfeeding as generally rather than very important had lower odds of belonging to the high self-efficacy group compared with the moderate group (odds ratio 0.18, 95 percent confidence interval 0.08 to 0.42), and fathers who expected non-exclusive feeding also had lower odds of high-group membership (odds ratio 0.38, 95 percent confidence interval 0.16 to 0.89). These associations remained unchanged after adjustment for paternal age, education, and household income, suggesting that the links between knowledge, attitudes, and self-efficacy profiles were not simply artifacts of demographic differences among the fathers studied. The authors frame the findings as a step toward early identification of fathers who may need additional support in the early postpartum period.

Because paternal support is described in the study background as a key determinant of breastfeeding success, with paternal self-efficacy serving as the cognitive foundation for fathers' involvement, distinguishing low-confidence subgroups could in principle allow clinicians to direct resources where they are most needed rather than applying uniform interventions. The researchers are careful to note the limits of what the study can establish. As a cross-sectional design with convenience sampling from a single hospital in Shanghai, it captures associations at one point in time and cannot demonstrate causal relationships, and the sample may not represent fathers more broadly.

The authors explicitly state that whether stratified support strategies built on these profiles actually improve paternal support behaviors or breastfeeding outcomes remains a hypothesis to be tested in future longitudinal and interventional studies. For now, the contribution is descriptive and methodological: a demonstration that paternal breastfeeding support self-efficacy is heterogeneous, and a preliminary map of the attitudes and knowledge levels that characterize fathers at the low and high ends of the confidence spectrum.

Subject of Research: Latent profiles of fathers' breastfeeding support self-efficacy in the early postpartum period

Article Title: A latent profile analysis of father support breastfeeding self-efficacy in the early postpartum period: a cross-sectional study

Article References: Li, J., Wang, Q., Pan, N., Zhang, L., Guo, N., & Liu, Y. (2026). A latent profile analysis of father support breastfeeding self-efficacy in the early postpartum period: a cross-sectional study. Frontiers in Nutrition, 13, 1951173. https://doi.org/10.3389/fnut.2026.1951173

Image Credits: AI Generated

DOI: 10.3389/fnut.2026.1951173

Keywords: breastfeeding, fathers, self-efficacy, latent profile analysis, postpartum period, paternal support, nutritional epidemiology, cross-sectional study, Shanghai, R3STEP procedure, breastfeeding knowledge, latent