Perinatal psychiatric symptoms and the continuum of breastfeeding in an urban university hospital population

Abstract

BACKGROUND

Breastfeeding has well-established public health benefits. We investigated the impact of perinatal psychiatric symptoms on reduced breastfeeding rates in an urban university community.

OBJECTIVE

This study investigates the relationship between perinatal psychiatric screening of symptoms on plans, initiation, and continuation of breastfeeding.

STUDY DESIGN

Pregnant women ≥18 years recruited from our urban academic clinic completed “My Mood Monitor (M-3)”, a validated psychiatric screening tool, and a breastfeeding questionnaire in 3 phases: second trimester (I- plans), postdelivery (II- initiation), and postpartum (III-continuation). Data were analyzed to establish the relationship between breastfeeding, maternal demographics, and the M-3 screening at these corresponding time points.

RESULTS

Although 86.3% (N=82) of our cohort planned to breastfeed, 76.1 % (N=54) initiated, and 43.4 % (N=23) continued breastfeeding. Early entry into prenatal care (≤12 weeks) was significantly associated with plans to breastfeed (*P=.02). At phase II, there was some association (⁺P=.065) between mothers with positive anxiety screening and breastfeeding, which became significant at continuation (*P=.014). Positive anxiety screening was significantly associated with postpartum discontinuation of breastfeeding (P=.019).

CONCLUSION

Early screening for identification and referral for diagnoses with treatment plans for psychiatric symptoms of perinatal anxiety, especially following delivery, may improve breastfeeding rates in urban minorities.

Introduction

Infant breastfeeding remains a cornerstone of child health and survival.^1,2 It provides essential and irreplaceable nutrition for growth, development, and protection from morbidities later in life.¹ It benefits both the mother-child dyad by improving bonding and contributing to successful emotional health.^1,3 This crucial preventive recommendation to breastfeed infants remains a major public health dilemma, with immense racial/ethnic disparities.^4,5 Despite an overall improvement in breastfeeding initiation rates at hospital discharge, non-Hispanic Blacks still trail behind non-Hispanic Whites.⁴ State-specific data from the District of Columbia reveal that breastfeeding initiation rates of non-Hispanic Blacks are 25.9 percentage points below those of non-Hispanic Whites and 11.3% below those of Hispanics.⁵ Addressing the Centers for Disease Control and Prevention Healthy People 2030 targets for the infants population of 42.4% to exclusively breastfeed and 54.1% to breastfeed up to 12 months old² requires immediate attention to minimize racial/ethnic gaps and to achieve these updated goals.

Studies identifying racial/ethnic barriers of breastfeeding intention to continuation,⁶ initiation⁷ and initiation to continuation,⁸ exist but adverse social determinants of health contribute to major disparities gaps in breastfeeding decision-making process and duration.⁹ In addition, the comorbid disparity of perinatal mood and anxiety disorders (PMADs)¹⁰ may derail successful breastfeeding decision.¹¹

The specific burden of mental health disorders in minoritized mothers has not been fully elucidated as a confounder of diminished perinatal breastfeeding goals. This study used a validated M-3 screen tool for a comprehensive group of perinatal psychiatric disorders (major depressive disorder [MDD], anxiety, bipolar disorder, posttraumatic stress disorder [PTSD], and suicidal ideation),¹² tracking longitudinally, to determine the relationship to breastfeeding achievements. From a public mental health perspective, it is pivotal to identify potential influences that comorbid psychiatric disorders have on the process of breastfeeding. This perinatal focus should aim to promote targeted intervention(s) to increase breastfeeding rates among minorities.

Methods

A longitudinal study was performed at Howard University Hospital in Washington, District of Columbia (DC), an urban institute with a minoritized patient population. The aim was to employ the My Mood Monitor (M-3), a screening tool for multiple psychiatric symptoms (MDD, bipolar disorder, anxiety, PTSD, and suicidal ideation), and to evaluate their relationship to breastfeeding at different time points. This study took place in its academic prenatal clinics and delivering hospital in urban Washington, DC, from October 2015 to May 2016. The Institutional Review Board at Howard University approved the study.

Study design and population

The enrollment criteria included pregnant women 18 to 45 years of age who attended the affiliated prenatal clinics and were willing to participate with a written informed consent, complete the study questionnaires, agree to have their medical records reviewed, and plan to give birth at Howard University Hospital’s labor and delivery suite. Non-English-speaking pregnant women were excluded; however, an inability to write in English was not an exclusion criterion. Those with no prenatal care and those who had known absolute contraindications for breastfeeding were excluded. The study team approached women in the prenatal clinic waiting area for private and confidential enrollment and surveys completion online or on paper, depending on the women’s choice. Strict HIPPA guidelines were adhered to.

Session structure and progression

The study was conducted in 3 phases (Figure 1).

Figure 1.

Flow diagram of cohort enrollment.

The study was conducted in these 3 phases of cohort enrollment. Each phase included the M-3 and breastfeeding questionnaires simultaneously.

Jain. Perinatal psychiatric symptoms and the continuum of breastfeeding. Am J Obstet Gynecol Glob Rep 2026.

Phase I-prenatal recruitment

Enrollment occurred during the second trimester of pregnancy (13–27 weeks gestation). The M-3 questionnaire was completed by the pregnant women, or it was read aloud by a research assistant when necessary. Response forms were deidentified. The phase I breastfeeding questionnaire, which assessed plans to breastfeed, was administered simultaneously with the phase I M-3 screen. To evaluate plans to breastfeed, responses from the questionnaire were used (answered YES to plan on feeding your baby breast milk only or plan on feeding your baby both breast milk and formula).

Phase II-postpartum

Women were followed longitudinally with the phase II breastfeeding questionnaire and phase II M-3 screening administered simultaneously, but once, either 2 days after delivery or before discharge from the hospital, when the mother was clinically stable. To evaluate the initiation of breastfeeding, responses from the phase II breastfeeding questionnaire were used (answered “YES” to breastfeeding was either initiated with breastmilk/breastfeeding alone or breastmilk/breastfeeding and formula feeding). Breastfeeding documentation from the electronic medical record was reviewed to corroborate the results from the questionnaire and confirm the initiation of breastfeeding.

Phase III-postdischarge

Postpartum women were rescreened after discharge. In conjunction with phase III M-3, the phase III breastfeeding questionnaire was repeated to evaluate the continuation of breastfeeding (answered “YES” to still putting your baby to your breast for feeding or putting your baby to breast and feeding your baby breast milk you pumped into a bottle or feeding your baby your breast milk that you pumped into a bottle only).

Measures

Psychiatric screen

The My Mood Monitor (M-3) is a validated screening tool for mental disorders.¹² It is a simple 27-question checklist designed for use in the primary care setting for adults aged 18 and older. The M-3 screens for MDD, bipolar disorder, anxiety disorder, PTSD, suicidal ideation, and measures functional impairment. The questionnaire (Q) is typically scored using a “gateway” approach, which requires that the patient’s answer be >0 on the suicide question or >1 on the sum of the functional impairment questions for the rest of the M-3 to be scored. If a gateway score is positive, scores are then calculated using a weighted algorithm for depression, anxiety, and bipolar screening. Patients who receive a positive screen for anxiety are then further examined for PTSD. Designated M-3 scoring above the threshold were as follows: MDD >5, Q 1 to 7; Bipolar >2, Q 20 to 23; Anxiety >3, Q 8 to 19; PTSD >2, Q 13 to 16; and suicide, Q 5. Surpassing these thresholds indicates that symptoms are commonly caused by a diagnosable mental health condition.¹³

The “gateway” approach was not used in this study, given prior findings that Blacks are less likely to endorse impairment in function, despite endorsing similar levels of symptoms for depression in epidemiological research.¹⁴ The recommended cutoff scores were used to screen for depression, anxiety disorder, PTSD, bipolar disorder, and suicidal ideation. All patients who screened positive were referred for further evaluation.

Breastfeeding measures

A breastfeeding questionnaire (Table 1) was developed to determine the participant’s intent to breastfeed (antenatal), initiation (after delivery), and continuation of breastfeeding postnatally. It correlated with the study time points. For the purposes of the study, breastfeeding was defined as: a) YES: both putting the baby on the breast to feed and/or providing expressed breast milk (that had been pumped) in a bottle to the baby. Mothers who did a combination of breastmilk/breastfeeding and formula feeding were categorized as breastfeeding. The exclusivity of breastfeeding was not evaluated separately. b) NO: no breastfeeding attempt as the baby received formula for feeding.

Table 1.

Phases of breastfeeding questionnaire

Phase I. Prenatal (plans to breastfeed)	Yes	No	Phase II- postdelivery (initiation of breastfeeding)	Yes	No	Phase III- postdischarge (continuation of breastfeeding)	Yes	No
1. Do you plan on feeding your baby BREAST MILK ONLY?			1. Have you ONLY breastfed your baby since birth			1. Are you STILL putting your baby to your breast for feeding?
2. Do you plan on feeding your baby BOTH BREAST MILK AND FORMULA			2. Up to 2 d following delivery, did you feed your baby both breast milk and bottle formula			2. Are you putting your baby to breast AND feeding your baby breast milk you pumped into a bottle?
						3. Are you feeding your baby your breast milk that you pumped into a bottle ONLY?

YES BREASTFEEDING – defined as both placing infant on the breast to feed and/or providing expressed breast milk (that had been pumped) in a bottle to the baby. Mothers who used a combination breastmilk/breastfeeding and formula feeding were categorized as breastfeeding. NO BREASTFEEDING - formula feeding infant only.

Jain. Perinatal psychiatric symptoms and the continuum of breastfeeding. Am J Obstet Gynecol Glob Rep 2026.

Statistical analysis

Data were analyzed using the Statistical Package for Social Sciences version 26. There were longitudinal timelines (phases I, II, and III) of cohort follow-up (Figure 1, Table 1). Patient demographics were cross tubulated with plans to breastfeed. Because our data are categorical, the χ² or Fisher’s exact test (for small samples) was used to establish significance of association where appropriate. All statistical tests were 2-sided, with significance established when the P-value was <.05. Descriptive statistics were conducted to assess the proportion of cohorts responding to breastfeeding during the established longitudinal times.

To establish the relationship of breastfeeding to any positive M-3 screen (MDD, anxiety, bipolar, PTSD, and suicidal ideation) at their corresponding time points, a similar cross-tabulation analysis was performed. This analysis was subsequently used to determine the associations of breastfeeding with specific M-3 screen disorders during the established longitudinal times.

Results

Demographic information

A total of 100 pregnant patients were approached to enroll in the study. Five women were eliminated due to incomplete initial assessment, leaving an N=95 for data analysis. Of the 95 women remaining for data analysis in phase I, the racial/ethnic makeup was 91 African American (96.8%), 3 Hispanics (3.2%), and 1 Asian (1.1%). The mean and median maternal age was 27 years, which was symmetrical, with a range of 18 to 44 years. The gestational ages of the infants ranged from 27 to 42 weeks (mean, 38 weeks). There was 1 triplet birth, and the others were singletons (Table 2).

Table 2.

Association between patient demographics with plan to breastfeed at phase I

Demographics	Plans to BreastfeedX (X/n%)(P-value)
Maternal age (years.): (N=95)  (Mean = Median)  <27 (n=46)  ≥27 (n=49)	40 (87.0%) 42 (85.7%) (P=.86)
Marital status: (N=61)  Single (n=47)  Married (n=14)	41 (87.2%) 13 (92.9%) (P=.56)
Gravida: (N=95)  Primigravida (n=35)  Multigravida (n=60)	32 (91.4%) 50 (83.3%) (P=.27)
Time of entry into prenatal care (weeks): (N=94)  ≤12 (n=42)  >12 (n=52)	40 (95.2%) 41 (78.8%) (*P=.02)
Number of prenatal visits: (N=95)  ≤4 (n=10)  >4 (n=85)	9 (90.0%) 72 (85.7%) (P=.86)
Pre-pregnancy BMI (kg/m2): (N=95)  <30 (n=61)  ≥30 (n=34)	55 (92.2%) 27 (79.4) (P=.14)
Gestational Age at delivery (weeks): (N=91)  <37 (n=14)  ≥37 (n=77)	13 (92.9%) 65 (84.4%) (P=.41)
Use of Recreational Drugs: (N=95)  No drug use (n=82)  Drug use (13)	12 (14.6%) 3 (23.1%) (P=.44)

Plan to breastfeed was separately cross tabulated with patient demographics at phase I. Note: Primigravida is defined as first-time pregnancy; Multigravida is defined as more than 1 pregnancy. An association between early entry into prenatal care (≤12 weeks) is noted to be significant *P=.02). None of the other patient demographics were significantly related to breastfeeding plans. N=Total number of maternal responses per group; X=Number of positive responses to breastfeeding questionnaire; n=demographic subset as defined. X/n % is the percentage of positive response for each defined subset.

Jain. Perinatal psychiatric symptoms and the continuum of breastfeeding. Am J Obstet Gynecol Glob Rep 2026.

Pregnancy-related data that were collected and corroborated with chart review include: maternal age, marital status, gravida status, the time of entry into prenatal care, number of prenatal visits, pre-pregnancy BMI, and gestational age at delivery. The use of recreational drugs (defined as tobacco, alcohol, specified illicit substance use, alone or in combination) was evaluated. Maternal demographics were assessed only in phase I, which included plans to breastfeed during pregnancy. There was a significant (*P=.02) association between early entry into prenatal care (≤12 weeks). Other patient demographics showed no relationship to breastfeeding plans (P>.05). The percentage use of recreational drugs during pregnancy was 15.8 %. We found no association between the decision to breastfeed during pregnancy and recreational drug use (P=.44). Additional data on domestic violence, unplanned pregnancy, and maternal education were challenging; hence, the small number of data sets collected could not be tabulated. Documented illness in our cohort (N=99) revealed none in 51% versus 48% with documented illness (either hypertension, diabetes, preterm labor, any mental illness, sickle cell disease, thyroid disease, or others), which did not show statistical significance for reporting. The frequency of mental illness by history or medications was 10%.

Those planning to initiate breastfeeding during pregnancy were N=82, 86.3%; at delivery, was N=54, 76.1 % (representing a higher actual initiation of breastfeeding than previously stated),⁵ and at continuation was N=23, 43.4 %. This frequency timeline represented a decline of 10.2% from planning to initiate breastfeeding to the time of delivery. In addition, there was another 32.7% decline in breastfeeding at postdelivery follow-up. Overall, the total decline in breastfeeding throughout the perinatal period was 42.9% (Table 3).

Table 3.

The frequency of planning to initiate breastfeeding during pregnancy was=86.3% (82/95); at delivery, it was 76.1 % (54/71) and 43.4 % (23/53) of the retained cohort continuing breastfeeding

	Positive M-3not breastfeeding	Positive M-3 and breastfeeding	Negative M-3not breastfeeding	Negative M-3 and breastfeeding
Plans to initiate N=95	9/13 (69.2%)	52/82 (63.4%)	4/13 (30.8%)	30/82 (36.6%)
Delivery N=71	2/17 (11.8%)	19/54 (35.2%)	15/17 (88.2%)	35/54 (64.8%)
Postpartum N=53	15/30 (50%)	4/23 (17.4%)	15/30 (50%)	19/23 (82.6%)

Cohort with positive M-3 and breastfeeding had a 28.2 % drop by from phase I to phase II (63.4% minus 35.2%). There was a further drop from phase II to phase III to 17.8% (35.2% minus 17.4 %). The overall decline in these cohorts from phase I to phase III was 45.6% (63.4% minus 17.8%).

Jain. Perinatal psychiatric symptoms and the continuum of breastfeeding. Am J Obstet Gynecol Glob Rep 2026.

We compare the relationship between the choice to breastfeed and a positive M-3 screening at their corresponding time points (Figure 2). Eighty percent of the cohort completed 2 M-3 questionnaires no earlier than 2 weeks apart, and the average of these 2 scores was used for data analysis. The remaining 20 percent completed 1 M-3 questionnaire. At the initial time point (phase I), those who had no plans to breastfeed with a positive M-3 screen (Blue bar) were represented by n=subset and N=total population; % population=9/13 (69.2%).This was simultaneously compared to the population with plans to breastfeed with a positive M-3 screen (Red bar), as 52/82=63.4%. At phase II, those who had no plans to breastfeed with a positive M-3 screen (Blue bar) were represented by n=subset and N=total population; % population=2/17 (11.8%). This was simultaneously compared to the population with plans to breastfeed with a positive M-3 screen (Red bar), as 19/54=35.2%. At phase III, those who had no plans to breastfeed with a positive M-3 screen (Blue bar) were represented by n=subset and N=total population; % population) as 15/30 (50%). This was simultaneously compared to the population with plans to breastfeed with a positive M-3 screen (Red bar), as 4/23=17.4%. In phase I, there was no significant association between plans to initiate breastfeeding (P=.684) and any positive M-3 screenings. At initiation, the percentage of breastfeeding decreased by 28.2% in mothers with any positive M-3 screening (⁺P=.065). At continuation, this percentage breastfeeding decreased by an additional 18% in mothers with any positive M-3 screening, showing a significant association (*P=.014). The overall decrease in breastfeeding with any positive M-3 screening from phase I to phase III was 46.2%.

Figure 2.

Comparing percentage responses of women with ANY M-3 screening over time.

Comparing percentage responses of women with ANY M-3 screening tabulated with their intent, initiation, and continuation of breastfeeding at corresponding time points. There was no significant association between plans to initiate breastfeeding (P=.684) and any positive M-3 screenings. At initiation, M-3 screen approached statistical significance (+P=.06). At continuation, M-3 screening, showed a significant association (*P=.014).

Jain. Perinatal psychiatric symptoms and the continuum of breastfeeding. Am J Obstet Gynecol Glob Rep 2026.

The relationship between breastfeeding in the 3 study phases (phase I, phase II, and phase III) and a positive M-3 specific subset (MDD, anxiety, bipolar, PTSD, and suicidal ideation) was examined. Cross tabulating the association between breastfeeding and positive M-3 subsets at corresponding time points revealed that the decrease in the continuation of breastfeeding (phase III) was associated with an anxiety screen (*P=.019). In addition, at this phase, there was a trend for an association with depressive disorder, but it was not significantly noted (⁺P=.058). No other statistical association with positive M-3 subset screening was noted (Table 4, Table 5, Table 6).

Table 4.

Plans to breastfeed and M-3 frequency (phase I)

Any positive M-3 (N=95)	Plans to breastfeed		Pearson X2=value, Asymptotic P-value (2-sided)
	Yes	No
Depression (n=43, 46.2%)	35 (81.4%)	8 (18.6%)	X2=1.42, P=.233
Anxiety (n=46, 48.4%)	40 (87%)	6 (13%)	X2=0.03, P=.86
Bipolar (n=32, 33.7%)	28 (87.5%)	4 (12.5%)	X2=0.06, P=.811
PTSD (n=21, 22.1%)	17 (81%)	4 (19%)	X2=0.66, P=.418
Suicide (n=20, 21.1%)	16 (80%)	4 (20%)	X2=0.86, P=.355

PTSD, posttraumatic stress disorder.

Total number of positive M-3 subset=n, % response. Depression (N=93).

Jain. Perinatal psychiatric symptoms and the continuum of breastfeeding. Am J Obstet Gynecol Glob Rep 2026.

Table 5.

Initiation of breastfeeding and M-3 frequency (phase II)

Any positive M-3 (N=71)	Initiation of breastfeeding		Pearson X2=value, Asymptotic P-value (2-sided)
	Yes	No
Depression (n=8, 11.3%)	7 (87.5%)	1 (12.5%)	X2=0.65, P=.421
Anxiety (n=9, 12.7%)	7 (77.8%)	2 (22.2%)	X2=0.02, P=.897
Bipolar (n=15, 21.1%)	13 (86.7%)	2 (13.3%)	X2=1.18, P=.278
PTSD (n=6, 8.5%)	5 (83.3%)	1 (16.7%)	X2=0.19, P=.662
Suicide (n=3, 4.2%)	3 (100%)	0 (0%)	X2=0.99, P=.321

PTSD, posttraumatic stress disorder.

Total number of positive M-3 subset=n, % response.

Jain. Perinatal psychiatric symptoms and the continuum of breastfeeding. Am J Obstet Gynecol Glob Rep 2026.

Table 6.

Breastfeeding after delivery and M-3 frequency (phase III)

Any positive M-3 (N=53)	Continuation of breastfeeding		Pearson X2=value, Asymptotic P-value (2-sided)
	Yes	No
Depression (n=11, 20.8%)	2 (18.2%)	9 (81.8%)	X2=3.59, +P=.058
Anxiety (n=13, 24.5%)	2 (15.4%)	11 (84.6%)	X2=5.5, *P=.019*
Bipolar (n=11, 20.8%)	3 (27.3%)	8 (72.7%)	X2=1.47, P=.225
PTSD (n=2, 3.8%)	2 (100%)	0 (0%)	X2=2.45, P=.118
Suicide (n=3, 5.7%)	1 (33.3%)	2 (66.7%)	X2=0.13, P=.717

PTSD, posttraumatic stress disorder.

Total number of positive M-3 subset=n, % response. PTSD (N=50).

Tables 4, 5, and 6: Cross tabulation of breastfeeding with the 5 specific positive M-3 screens at corresponding time points showed that there was a significant association (χ²=5.5; *P=.02) with continuation of breastfeeding with positive M-3 screen for anxiety screen after delivery (phase III). Depression screen did not meet statistical significance (χ²=3.59 ⁺P=.06) in phase III.

Jain. Perinatal psychiatric symptoms and the continuum of breastfeeding. Am J Obstet Gynecol Glob Rep 2026.

Discussion

The population in the District of Columbia shows a high incidence of health disparities, along with challenges in early and regular entry into prenatal care that could improve pregnancy and birth outcomes.¹⁵ This unique study involving this urbanized population showed that early entry into prenatal care (≤12 weeks) is significantly associated with plans to breastfeed (*P=.02); however, sustaining breastfeeding preference declined by 42.9% over time.

There was a significant challenge with postdischarge breastfeeding continuation (P=.014) as screening symptoms of anxiety significantly contributed to breastfeeding discontinuation (P=.019). The reported prevalence of anxiety disorders during pregnancy and the early postpartum period (15.8% and 17.1%, respectively), exceeding that of depression (3.9% and 4.8%, respectively).¹⁶ Our study showed incidences of anxiety of 48.4% during pregnancy, 12.5% after delivery and 24.5% following delivery. Anxiety can also be a pre-existing condition that worsens during the postpartum period.¹⁷ Some studies show that prenatal anxiety and depressive disorders predicted lower intents to breastfeed after adjusting for demographics,¹⁸ including Hispanic mothers.¹⁰ Our study showed a concurrence of depressive symptoms following delivery; however, this did not reach statistical significance.

The study cohort had higher rates of intention to breastfeed than other studies involving minoritized women.19, 20, 21 Intending to breastfeed antenatally has been shown to be strongly associated with good breastfeeding outcomes at discharge,¹⁹ even in women with severe²¹ and without mental health disorders.¹⁹ Our cohort showed a slightly higher initiation rate of breastfeeding (76.1%) compared to current national data for Black mothers (73.6%) and higher than the initiation rates in the District of Columbia (71.2%).⁵

Following delivery (phase II), mothers were engaged in breastfeeding despite having positive M-3 screens. This could have been an opportunity to improve breastfeeding rates while in the hospital, to perhaps sustain more breastfeeding opportunities following discharge, using the Baby-Friendly Hospital Initiatives (BFHI).²² However, our hospital site has not yet incorporated BFHI.

In our data, the continuation of breastfeeding following discharge in mothers with any positive M-3 screening (of combined symptoms) showed an additional significant 17.8% decline from the delivery time (see Table 3). Studies support maternal postpartum anxiety as a risk factor for not achieving exclusive²³ and prolonged breastfeeding.^23,24 With regard to racial and ethnic differences, Non-Hispanic Blacks have a higher risk of combined perinatal depression, comorbid depression/anxiety, and moderate/severe to severe depression.²⁵ This study supports the published literature on the impact of postdischarge anxiety on the early cessation of breastfeeding and suggests that screening and interventions targeted only towards depressive symptoms alone may not be sufficient to improve breastfeeding rates. The shift beyond just MDD screening²⁶ must now include a heterogeneous group of PMADs^3,25,27 as comorbid conditions.^13,14,16 Thereafter, the mechanism through which perinatal anxiety impedes a woman’s ability to breastfeed following discharge warrants further study in minority groups.

Among common comorbid psychiatric disorders with high racial/ethnic disparities in the United States rates, MDD has been the most studied for its effect on breastfeeding practices, with a 10% to 30% prevalence antenatally and an estimated postpartum prevalence of up to 40% to 60% among low-income mothers.²⁸ Our study did not show associations with other PMAD while using the M-3 screen.

The strength of this study provides a description of prospective data; it was not subjected to major recall bias. In addition, it was done on a population that has been underrepresented in many studies. Using the M-3 tool allows for a single, broader, quicker, and easy-to-use screening of other comorbid mental health conditions that meet the current American College of Obstetricians and Gynecologists Clinical Practice Guidelines.²⁹ The current clinical use of the Edinburgh Postnatal Depression Scale (EPDS) has limited mental health screening^26,30 omitting recommended psychiatric conditions when used alone, which will guide diagnosis and psychopharmacotherapy as indicated, to optimize perinatal health.

This study has limitations inherent to self-reporting questionnaires, which tend to be more sensitive than specific. There is the possibility of bias from participants reporting their feeding choices in a manner that they perceived to be socially desirable. Our sample size was small, with limited geographic and socioeconomic representation; hence, the results may not be applicable to a general population of all perinatal minoritized women. The exclusivity of breastfeeding was not a focus in this study. Although we targeted a 6-week postdischarge follow-up, we had an average of 2.3 months (range: 1–7 months) of follow-up due to losses and logistical issues.

We used a newer screening tool to reveal multiple comorbid psychiatric conditions of pregnancy. It is not a diagnostic tool with clinical interviews, nor was information systematically collected on whether women received psychiatric intervention following referral. The attrition rates were high from phase I to phase III; it was possible that these women may not have followed up with the recommended psychiatric referral and may have had ongoing, unmanaged perinatal psychiatric disorders, making them less likely to breastfeed. Information was not collected on prior experience with breastfeeding in phase I, nor were additional reasons for discontinuation of breastfeeding (medical and non-medical) collected at phase III. Nevertheless, our results identify postpartum anxiety as a significant contributor to early breastfeeding cessation.

A study done by Spinelli et al,³¹ observed that a higher than usual rate of breastfeeding in American-born Black women can be modified by educational support and intervention for their depressive ailments, regardless of their marital or socioeconomic status. Further research is warranted to determine the optimal time for PMAD screening and intervention in mothers with a positive intent to breastfeed during pregnancy. Interventions that support the transition from high breastfeeding intent to actual initiation should be prioritized to achieve the desired breastfeeding outcomes.

Conclusion

Our pilot study shows the challenges that minoritized urban pregnant women have in the continuation of breastfeeding. A timely and early screening by obstetric providers at pregnancy visits should incorporate a standardized, evidence-based comprehensive screening beyond that of the EPDS screen. Positive screening for mental health condition(s) needs referral to psychiatry for diagnosis, management, and follow-up to improve breastfeeding rates following delivery. These important observations can be utilized for future research along with public health strategies for perinatal interventions to support improved breastfeeding practices in underserved populations.

CRediT authorship contribution statement

Swati Jain: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Data curation, Conceptualization. Tomilowo Abijo: Writing – review & editing, Validation, Formal analysis, Data curation, Conceptualization. Edwin W. Powell: Supervision, Software, Methodology, Investigation, Data curation, Conceptualization. Sindile Dlamini: Writing – original draft, Resources, Methodology, Conceptualization. Ethan Kellogg: Writing – original draft, Visualization, Validation, Resources, Methodology, Investigation, Data curation, Conceptualization. Inez Reeves: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.