Breastfeeding Behaviors and Maternal Interaction Quality in a Low-Income, Ethnic Minority Population

Abstract

Objective

To examine the associations between breastfeeding intensity and underexplored features of maternal-child interaction quality, over and above the influence of breastfeeding initiation.

Method

The current study leveraged an on-going, multi-site randomized controlled trial of a tiered parenting program for 462 Medicaid-eligible mothers and their infants in the United States. We examined whether breastfeeding intensity and exclusivity was associated with observed maternal sensitivity, intrusiveness, and detachment, as well as self-reported maternal verbal responsiveness, at infant age 6 months. Analyses controlled for breastfeeding initiation, demographics, and early parenting experiences.

Results

Higher intensity breastfeeding at 6 months significantly was related to higher maternal sensitivity (β=0.12, p=0.004) and lower maternal intrusiveness (β=−0.10, p=0.045). There was no significant association between breastfeeding intensity at 6 months and detachment (β=−0.02, ns) or self-reported verbal responsiveness (β=0.11, ns). Results were the same when intensity was measured as a dichotomous indicator for exclusive breastfeeding. Effect sizes were small-to-moderate, ranging from Cohen’s d=0.26 to 0.31. Associations did not vary by site, race/ethnicity, infant difficultness, or household poverty.

Conclusion

The finding that breastfeeding intensity was significantly and independently associated with maternal sensitivity and intrusiveness is novel in the literature on low-income families from the United States. These findings have implications for breastfeeding promotion strategies and indicate that future research should explore synergistic or spillover effects of interventions aimed at maternal-child interaction quality into the infant feeding domain, particularly in the primary care setting.

INTRODUCTION

Breastfeeding is a primary feature of many mother-infant relationships, often involving sustained and regular dyadic interaction for weeks, months, or even years. While substantial research has focused on the physiological and immunological benefits of breast milk for the infant, a growing body of literature focuses on the psychological experience of breastfeeding for the dyad.¹ Breastfeeding is associated with increased maternal sensitivity and more secure mother-infant attachments.^2,3,4,5,6,7 One hypothesized mechanism is that breastfeeding mothers smile, talk, and touch their infants more frequently during feeding interactions when compared with non-breastfeeding mothers.⁸ Perhaps synergistically, breastfeeding is related to greater activation in maternal brain regions that are implicated in maternal-infant bonding.⁹ Building on this work exploring the dyadic relationship during feeding, researchers are exploring whether and how breastfeeding may relate to other relational features of the mother-infant relationship.

One limitation of studies to date has been their focus on only breastfeeding initiation^1,2,3,4 or the duration of any breastfeeding.^5,6,7 Breastfeeding behaviors exist on a continuum across a range of breastfeeding intensities, from token breastfeeding (e.g., once-per-day) to on-demand, exclusive breastfeeding (e.g., 8 or more times per day, every day, with no formula feeding).¹⁰ A number of studies have found a relationship between more breastfeeding and positive child outcomes (e.g., reduced risk of ear infections¹¹ and higher adolescent IQ¹²). Analyses of the association between breastfeeding intensity and/or exclusivity and mother-child interaction quality are more limited. The critical notion here is that of a dose-response relationship between a behavior and an outcome. By dose-response, we mean a greater benefit resulting from a higher “dose” of breastfeeding than from a lower “dose” of breastfeeding. One aspect of breastfeeding dosage is duration of any breastfeeding, which has been examined; the other aspect is breastfeeding intensity and/or breastfeeding exclusivity (i.e., the proportion of breastmilk vs. infant formula provided to the infant), which has been neglected in the literature. In general, findings of a dose-response relationship would suggest that the dose is causally influencing the response.¹³ Uncovering a dose-response relationship between breastfeeding behaviors and mother-child interaction characteristics therefore necessarily involves understanding the role of both duration and intensity/exclusivity play in these dyadic outcomes.

Although the correlations between infant feeding methods and various aspects of mother-child interaction quality are established, the theoretical directionality of those relationships is unclear. For instance, it may be that more sensitive or responsive mothers have stronger intentions to breastfeed or are more successful at meeting those breastfeeding intentions.² Alternatively, it may be that the act of breastfeeding itself boosts maternal sensitivity or provides mothers with more practice in sensitively responding to their infants, which then spills over to other aspects of the mother-child relationship.^3,5,9 More likely, both of these pathways are operating simultaneously, resulting in a transactional relationship that cascades forward.⁷ Regardless of the causal pathways, basic unanswered questions remain regarding which specific dimension of the breastfeeding relationship are related to specific features of the mother-child interactions (i.e., initiation vs. duration vs. intensity). Furthermore, understanding the relationship between each dimension and maternal interaction quality may inform the discussion around directionality. For instance, if breastfeeding initiation is related to positive mother-child interaction qualities, but sustained, high-intensity breastfeeding is not, that might suggest some third, unmeasured covariate is driving changes in both interaction quality and feeding method. If, however, sustained breastfeeding is related to interaction quality over and above the influence of breastfeeding initiation, this suggests that breastfeeding itself may confer some benefit to the development of positive parenting.

Another limitation of the existing studies linking breastfeeding to the dyadic relationship is that the majority of studies have taken place either outside of the United States or in middle- to high-income U.S. communities, potentially limiting their generalizability to low-income, Medicaid-insured, and/or ethnically and racially diverse United States populations at greater risk of health disparities.^2,4,6,7,14 Based on the lower rates of exclusive breastfeeding and shorter breastfeeding duration in low-income populations, lower income groups deserve more focused attention.³

To fill these gaps, we sought to determine the associations between breastfeeding intensity and exclusivity and maternal interaction quality, utilizing a multi-method approach to capture the later, with both observed videotaped play interaction at 6 months old and maternal self-reported verbal responsiveness. This study was conducted with a U.S. sample spanning two cities with a primarily African-American and Latina population of Medicaid-eligible mothers. We hypothesized that breastfeeding initiation would have a positive association with indicators of high-quality interactions (e.g., highly sensitive or responsive parenting interactions), and that initiation would have a negative association with indicators of low-quality interactions (e.g., intrusive or detached parenting interactions). Although there is a dearth of research on the relationship between parenting and breastfeeding beyond initiation, we hypothesized that greater breastfeeding intensity and exclusivity would also be related to more sensitive and less intrusive and detached interactions.

METHODS

Study Design

We conducted a secondary analysis of observational data collected as part of an ongoing study of the Smart Beginnings model. Smart Beginnings integrates two evidence-based interventions: 1) a universal primary prevention strategy (Video Interaction Project [VIP]); and 2) a targeted secondary prevention strategy (Family Check-up [FCU]) for families identified as having additional risks. The VIP infant-toddler program consists of 14 sessions that take place between birth and three years, scheduled at the time of well-child appointments, occurring more often in the first year (6 times) and then less frequently in years 2 and 3 (every three months). VIP provides parents with a developmental specialist who videotapes the parent and child and coaches the parent on effective parenting practices at each pediatric primary care well-child visit.¹⁵ FCU is a home-based, family-centered intervention that utilizes an initial ecologically-focused assessment to promote motivation for parents to change child-rearing behaviors, with individualized follow-up sessions on parenting and factors that compromise parenting quality beginning at 6 months.¹⁶

Following informed consent procedures, mother-infant pairs were enrolled in the newborn nursery. A first cohort was randomly assigned to 2 treatment conditions: the Smart Beginnings intervention condition, which received VIP and potentially also received FCU, or the control group, which receives only routine primary care (n=405). A second cohort was enrolled to VIP only but enrolled under the same inclusion criteria (n=57). The present secondary analysis included participants from both cohorts. Breastfeeding initiation, intensity, and exclusivity were assessed using survey measures at infant age 6 months old. Verbal responsiveness was also assessed by self-report survey at 6 months. Maternal interaction quality was assessed during a videotaped play interaction at 6 months old, following random assignment to Smart Beginnings and receipt of 4 VIP sessions (1 month, 2 month, 4 month, and 6 month), but before any FCU intervention had taken place. The institutional review boards of the hospitals in which the trials were conducted approved this study.

Study Sample

Participants came from hospital-based clinics serving low-income communities in 2 urban centers in the Northeast: New York City (site 1), and Pittsburgh (site 2). Site 1 sample consisted of 257 participants, and the Site 2 sample consisted of 205 participants. The study included mothers with infants who intended to receive pediatric care at the study sites for the next three years, whose primary language was English or Spanish, and who reported having a working phone. We excluded mother-infant pairs with infant birth weight <2500gm, gestational age <37 weeks, and with known or suspected significant genetic abnormality, neurodevelopmental or neuromuscular disorder, sensory defect, significant malformation likely to affect development or likely to require significant therapy, postnatal complication requiring level II or III nursery stay (e.g., sepsis, significant hypoglycemia, seizures), meeting criteria for Early Intervention at birth, or concerns about infant hearing or vision. We also excluded women with multiple gestation pregnancy, with known significant impairment that would be a barrier to communication and participation (e.g., intellectual disability, schizophrenia), were living in shelter at the time of enrollment, or if their infant was not discharged to them or the father. Mothers provided written informed consent prior to enrollment, and once enrolled, were administered a baseline assessment before infant age 6 weeks. All participating families completed the same assessments at the same time points, and were recruited using the same procedures and inclusion / exclusion criteria. The analysis sample included mother-infant pairs with complete data at baseline and 6-month assessments from both cohorts (n=396).

Assessments

Mother-Child Interaction Quality

Our first outcomes of interest were global measures of mother-child interaction quality during a 5-minute videotaped, free-play task based on the Parent-Child Interaction Rating Scales-Infant Adaptation (PCIRS-IA). The PCIRS-IA is an observational scale adapted from three separate scales to assess the quality of the primary caregiver’s parenting behaviors and their interactions with the target child: The Parent-Child Early Relational Assessment (ERA)¹⁷; the Mother-Child Interaction Rating Scale (MCIRS)¹⁸; and the Caregiver-Child Affect, Responsiveness, and Engagement Scales (C-Cares).¹⁹ The PCIRS-IA is made up of mother ratings (8) and dyadic ratings (3). We utilized three of the ratings that mapped most closely to the various aspects of maternal sensitivity and responsiveness: sensitivity, intrusiveness, and detachment. Coders were blind to random assignment status and all other information about the mother and infant, including infant feeding methods. Each item was scored on a scale from 1 (Very low), meaning the mother did not demonstrate the described behavior, to 7 (Very high) meaning the mother characteristically demonstrated the described behavior. The average weighted reliability kappa for these ratings ranged from 0.70–0.85, indicating a high degree of agreement between coders.

Our second outcome of interest was self-reported maternal verbal responsiveness, as measured by the Infant StimQ (StimQ-I) Parental Verbal Responsivity subscale (PVR).²⁰ The StimQ-I is a validated scale of cognitive stimulation in a home environment of infants provided by the primary caregiver, including language quality, book reading, and activities. Specifically, the PVR subscale consists of 13 yes/no items about verbal interaction quality during everyday routines and responsiveness during pretend play. The PVR is scored by summing the number of affirmative (yes) responses to each item (sample α = 0.65).

Breastfeeding

Our independent variables were three features of the breastfeeding relationship: a dichotomous indicator for breastfeeding initiation (did mother ever breastfeed), and two measures of breastfeeding intensity at 6 months. Intensity was first operationalized on a 0–4 scale: 0=Exclusive formula feeding (no breastmilk), 1=More formula than breastmilk, 2=About the same amount of formula and breastmilk, 3=More breastmilk than formula, 4=Exclusive breastfeeding (no formula). We also utilized a dichotomous indicator for breastfeeding exclusivity at 6 months (only breastfeeding, no formula vs. all others at 6 months).

Covariates and Moderators

Several maternal and infant characteristics were crucial to account for in exploring these relationships because they were associated with maternal interaction quality, infant feeding method, or both. Because this trial was conducted at two sites, we controlled for location (indicator for site fixed effects) and treatment assignment (indicator for assignment to VIP vs. control) in all analyses. We also controlled for maternal socio-demographic factors, including parity (dichotomized for first child), age (dichotomized for teen motherhood, under age 20 at child birth), education level (categorical), marital status (dichotomized for legally married or not), and race / ethnicity indicators. Other maternal characteristics were self-reported in the baseline survey and included baseline level of maternal depressive symptoms (Edinburgh Postnatal Depression Scale (EPDS)²¹ total score at birth; α = 0.87) and maternal life satisfaction and social support (General Life Satisfaction (GLS)²² total score at birth; α = 0.78). Finally, we controlled for infant temperamental difficultness (Infant Characteristics Questionnaire (ICQ) Difficultness²³ subscale score which was parent-reported at 6 months; α = 0.79) and infant age at 6-month assessment (continuous).

Moderation analyses were also conducted using income-to-needs ratio and food insecurity. Income-to-needs is a standard measure of a family’s economic situation, which represents the number of individuals supported by the family income.²⁴ The ratio divides a family’s total income by its corresponding poverty threshold, or the dollar amount used to determine poverty status using size of the family and number of related children under 18 living in the house. An income-to-needs ratio of 1.00 indicates that a family’s income is exactly at the poverty threshold for that size family; 2.00 indicates that a family is 200% above that threshold. Food insecurity was measured using a shortened version of the Household Food Security Survey (HFSS).²⁵ The HFSS is an 18-item survey developed by the U.S. Department of Agriculture to identify families at risk for food insecurity, adapted by Hager and colleagues into a validated 2-item screen that has proved sensitive and valid among low-income families with young children.²⁶ Respondents were asked to respond Often true, Sometimes true, or Never true to the following questions: “Within the past 12 months we worried whether our food would run out before we got money to buy more” and “Within the past 12 months the food we bought just didn’t last and we didn’t have money to get more.” Higher scores indicate more food insecurity.

Statistical Analysis

Data analyses were performed using Stata version 14.0. All associations were tested using ordinary least squares (OLS) linear regression controlling for each covariate listed above. For dichotomous variables (breastfeeding initiation and exclusive breastfeeding at 6 months), effect size was estimated using Cohen’s d, which is calculated by dividing the difference in adjusted means between the two groups by the pooled standard deviation.²⁷ We tested the association between breastfeeding initiation and the four aspects of mother-child interaction quality -- sensitivity, intrusiveness, detachment, and verbal responsiveness. Then we tested the association between continuously measured breastfeeding intensity and mother-child interaction quality, controlling for breastfeeding initiation. Next, we tested the association between exclusive breastfeeding at 6 months and mother-child interaction quality, controlling for breastfeeding initiation.

We conducted several sensitivity analyses. First, we ran each regression by site to determine whether the associations were site-specific. Next, we tested whether any of the associations between breastfeeding intensity and the interaction quality measures were significantly moderated by maternal race, income-to-needs ratio, food insecurity, or infant difficultness (using the ICQ Difficultness scale). This amounted to including in the regressions interaction terms between breastfeeding intensity and exclusive breastfeeding, and each moderator. Finally, we dropped participants who were not randomly assigned (n=56) from the analysis sample to test if this modeling specification altered the results.

RESULTS

Study Sample

We enrolled and consented 462 mother-infant dyads, all of whom completed baseline assessments. Of the 462 enrolled participants, 409 (89%) completed the baseline and 6-month survey assessment including breastfeeding data, and 396 of the 462 (86%) completed all observational videotaped play interactions in addition to the surveys. Of those who completed the survey measures, the majority self-reported as African American (40.6%) or Hispanic (49.9%). 31.6% had less than high school education, 21.5% were married, and 10.6% reported moderate depressive symptoms (Table 1). 81% of mothers reported breastfeeding initiation. At infant age 6 months, mothers reported the following feeding patterns: 53% no breastmilk (exclusive formula feeding), 13% more formula than breastmilk or the same amount of formula and breastmilk (combination feeding), 9% more breastmilk than formula (predominantly breastfeeding), and 26% exclusive breastfeeding.

Table 1:

Sample Demographic Characteristics and 6-month Breastfeeding Intensity by Characteristic, n=409

	Total Sample (%)	No Breastfeeding N=216 (%)	Combination Feeding N=53 (%)	Predominant Breastfeeding N=35 (%)	Exclusive Breastfeeding N=105 (%)	X2 Difference by Characteristic
Maternal Age						2.3
<20	5.3	66.7	14.3	4.8	14.3
20+	94.7	52.1	12.9	8.8	26.3
Education						34.2***
Less than HS	31.6	33.3	18.6	13.2	34.9
HS Diploma / GED	33.5	64.5	13	6.5	15.9
More than Diploma	34.9	59.2	7.8	6.3	26.8
Race / Ethnicity						101.5***
Af. Amer. / Black	40.6	77.3	4.9	4.3	13.5
Latina / Hispanic	49.9	30.5	21.7	13.8	34
White	5.4	81.8	0	0	18.2
Asian Amer. / PI	1	0	25	0	75
Other	3.2	61.5	0	0	38.5
Marital Status						18.5***
Married	21.5	34.5	18.4	16.1	31
Not Married	78.5	58.4	11.4	6.6	23.7
Depressive Symptoms
EPDS < 10	10.6	52.1	13.7	9	25.2	3
EPDS 10+	89.4	59.1	6.8	4.6	29.6
Infant Temperament
ICQ_D < 20	68.2	50.0	12.1	9.6	28.2	5.2
ICQ_D 20+	31.8	58.9	14.7	6.2	20.2

^***p<0.001

Note: HS=High School; GED=General Education Diploma, or High School Equivalency Exam; PI=Pacific Islander; EPDS=Edinburgh Postnatal Depression Scale; ICQ-D=Infant Characteristics Scale, Difficultness Subscale

In bivariate analyses, mothers who were exclusively breastfeeding at 6 months were more likely to be Hispanic than those who were not exclusively breastfeeding (χ² = 101.5, p < .001). Participants who had completed high school but did not have additional education were less likely to exclusively breast feed compared to those who had not completed high school or to those who had more than a high school education. There were no associations between breastfeeding intensity and either maternal depressive symptoms or infant difficultness (Table 1).

Breastfeeding Initiation and Maternal Interaction Quality

Breastfeeding initiation was associated with lower maternal intrusiveness (β=−0.45, p=0.026). There was no significant association between breastfeeding initiation and detachment (β=0.02, ns), maternal sensitivity (β=0.29, ns), or self-reported verbal responsiveness (β=−0.11, ns).

Breastfeeding Intensity and Maternal Interaction Quality

See Table 2 for results. Breastfeeding intensity at 6 months significantly was associated with higher maternal sensitivity (β=0.12, p=0.004) and lower maternal intrusiveness (β=−0.10, p=0.045) after accounting for the influence of breastfeeding initiation. There was no significant association between breastfeeding intensity at 6 months and detachment (β=−0.02, ns) or self-reported verbal responsiveness (β=0.11, ns) after controlling for the influence of breastfeeding initiation.

Table 2:

Predicting Interaction Quality from Breastfeeding Intensity and Exclusive Breastfeeding at 6 Months, n=396

	Coef.	Std. Err.	95% CI		P-Value
Breastfeeding Intensity
Sensitivity	0.12	0.04	0.04	0.20	0.004
Detachment	−0.02	0.02	−0.06	0.02	0.308
Intrusiveness	−0.10	0.05	−0.20	0.00	0.045
Verbal Responsiveness	0.11	0.10	−0.08	0.30	0.270
Exclusive Breastfeeding
Sensitivity	0.37	0.15	0.08	0.66	0.012
Detachment	−0.05	0.08	−0.20	0.10	0.501
Intrusiveness	−0.37	0.18	−0.72	−0.03	0.035
Verbal Responsiveness	0.48	0.35	−0.22	1.16	0.177

Note: In addition to breastfeeding initiation, all analyses controlled for parity, education, marital status, teen motherhood, intervention site, treatment assignment, race/ethnicity, maternal life satisfaction and depression, child age, and child temperament.

Breastfeeding Exclusivity and Maternal Interaction Quality

See Table 2 for results. Breastfeeding exclusivity at 6 months significantly was associated with higher maternal sensitivity (β=0.37, p=0.013) and lower maternal intrusiveness (β=−0.37, p=0.034) after accounting for the effect of breastfeeding initiation (Table 2). There was no significant association between breastfeeding exclusivity at 6 months and detachment (β=−0.05, ns) or self-reported verbal responsiveness (β=0.47, ns) after controlling for breastfeeding initiation. The effect sizes for the significant associations between exclusive breastfeeding and maternal sensitivity and intrusiveness were small-to-moderate, ranging from 0.26 to 0.31 (Figure 1).

Figure 1:

Effect Size of Initiation of Breastfeeding and Exclusive Breastfeeding at 6 Months on Maternal Interaction Quality, n=396

*p<0.05

Note: Effect sizes calculated using Cohen’s D with pooled standard deviations.

Sensitivity Analyses

When the analysis sample is split by site (n=219 in Site 1, n=177 in Site 2), the direction and magnitude of associations were very similar across site. However, the associations do not reach statistical significance because they are underpowered. We also conducted sensitivity analyses to test whether the associations between breastfeeding intensity and the maternal-infant interaction quality measures were moderated by income-to-needs, food insecurity, by race/ethnicity, or infant temperament. None of the associations varied by any of these measures. Tables of those results are available upon request.

Finally, we conducted the analyses with and without those participants who were assigned to a VIP-only (non-randomized) condition (n=56). There were no change in findings, except for reduced power because of a smaller sample size. Tables of those results are also available upon request.

DISCUSSION

In this sample of Medicaid-insured, low-income women and their infants, we found that more intensive breastfeeding at 6 months was associated with higher sensitivity and lower intrusiveness after accounting for the influence of breastfeeding initiation. We also found that breastfeeding initiation marginally related to lower observed maternal intrusiveness in a videotaped play interaction at 6 months. We did not find significant associations between indices of breastfeeding intensity and observed maternal detachment or self-reported verbal responsiveness.

The association found between higher “dosage” of breastfeeding and higher maternal sensitivity / lower maternal intrusiveness is consistent with other recent research that mothers who choose to breastfeed are also more likely to be sensitive in responding to their children. This validates those findings for a low-income, ethnic-minority population.^5,14 We extended the literature by examining how the quality of mother-child interaction is associated with breastfeeding intensity and exclusivity, over and above the effect of breastfeeding initiation, infant difficultness, and maternal postnatal depression and life satisfaction. We found that these associations did not differ by income-to-needs ratio, food insecurity, infant difficultness, or maternal race, and were instead relatively homogenous across the diverse sample. The results were also very similar across the two urban centers. The finding that breastfeeding intensity is significantly and independently associated with maternal sensitivity and intrusiveness is novel in the literature on low-income families.

A variety of mechanisms may be responsible for the association between breastfeeding and positive parenting interactions. Oxytocin, which is released in large doses during a breastfeeding session, is a hormone that influences the maternal neurocognitive circuitry that is involved in processing infant distress and other emotional and social stimuli, and facilitates social bonding and affectionate parenting behavior.²⁸ Breastfeeding mothers exposed to their own infant’s crying showed greater activation in brain regions known to be associated with infant bonding from animal studies, such as the superior frontal gyrus and amygdala, compared to exclusively formula-feeding mothers.⁹ In addition to biologic changes, breastfeeding also represents periods of close physical contact, more touching, and affectionate interactions between the infant and mother.¹ Close physical contact has been demonstrated to protect against maternal stress, which may ultimately contribute to increased sensitivity to infant cues. The more mothers breastfeed, the more they may become aware of, and appropriately responsive to, their infant’s needs over time.^1,7

These findings support that the benefits of breastfeeding may extend beyond improving maternal and child health outcomes, enhancing parenting more generally and improving the mother-infant relationship. While many primary care-based breastfeeding promotion programs have shown success in increasing initiation, duration, and exclusivity of breastfeeding, few studies have explored their effects on the quality of the maternal-infant relationship.²⁹ Likewise, many primary care-based interventions aiming to promote positive parenting have been successful at increasing parental sensitivity to child needs.³⁰ However, very few studies have examined spillover from these types of programs. In other words, how might interventions in the primary care setting that directly target either breastfeeding or sensitive parenting affect the other, non-targeted domain? More research is needed to explore how these two processes overlap, and to leverage spillover effects or to combine interventions for maximal cost-effectiveness.

This study has several limitations. First, given that breastfeeding initiation was assessed at infant age 6 months and not around the time of birth, there may be recall or social desirability bias present in that mothers may have reported initiating breastfeeding when they have not done so. Second, though we accounted for several demographic and psychosocial differences between breastfeeding and non-breastfeeding mothers in our analyses, including infant temperamental difficultness, maternal life satisfaction, and depressive symptoms, there may be unobserved factors that explain observed differences in parenting other than infant feeding behaviors. Third, sensitivity analyses showed that results were similar, but not identical, when non-randomized participants were included in the analysis, despite controlling for assignment to treatment condition. This may be due to underpowered analyses with the smaller analytic sample, or due to differences between the two cohorts related to unmeasured, confounding variables. Further, while our study aimed to explore observed maternal detachment, rates of these parenting qualities were quite low in this sample and likely hindered the power to detect any associations with breastfeeding. Additionally, though the magnitude of the association between breastfeeding intensity and the parent-report measure of verbal interaction quality was similar to the association with maternal sensitivity and intrusiveness, those associations did not reach statistical significance. While this may be due to lower than expected reliability of the StimQ measure in this particular sample, it is also possible that breastfeeding is more related to quality of interactions rather than presence and quantity of specific cognitive stimulation activities that are the focus of the StimQ. Finally, this study was one of the first to focus on low-income Hispanic and African American families in diverse urban settings, which is a major strength of the study. However, findings may not be generalizable to other racial, ethnic, or geographic groups.

In summary, this study demonstrated moderate associations between higher breastfeeding intensity and higher quality maternal interactions during play in low-income maternal-infant pairs. Longitudinal and causal studies are needed to determine whether and how these features of the family are related over time. These studies would help to inform how efforts to strengthen and expand existing breastfeeding promotion programs could further promote sensitive parenting in high-risk communities.