SOCIAL DETERMINANTS OF BREASTFEEDING: THE ROLE OF PRENATAL FOOD INSECURITY By Chelsea Robinson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Epidemiology—Master of Science 2022 ABSTRACT SOCIAL DETERMINANTS OF BREASTFEEDING: THE ROLE OF PRENATAL FOOD INSECURITY By Chelsea Robinson Background: Relatively little work has quantified associations between prenatal food insecurity and breastfeeding practices; however, understanding the implications of prenatal food insecurity may support food insecurity screening recommendations during prenatal care. Therefore, the purpose of this study was to investigate associations between prenatal food insecurity and breastfeeding initiation and duration. Method: This study utilized data from a prospective Michigan pregnancy cohort. Women were recruited during their first prenatal visit. Prenatal food insecurity was assessed during pregnancy, and breastfeeding initiation and duration were assessed at the 3-month postpartum visit. Multiple logistic regression models were used to evaluate associations between prenatal food insecurity and two primary outcomes: breastfeeding initiation and breastfeeding status at 3-months postpartum. Cox proportional hazard ratios were used to assess differences in the risk of breastfeeding cessation until 3 months postpartum by food insecurity status. An adversity index was created to stratify women into higher- and lower-risk groups for not breastfeeding. Associations between food insecurity and breastfeeding at 3 months postpartum were assessed via Fisher’s Exact test within each group. Results: In the unadjusted models, women who reported prenatal food insecurity were less likely to initiate breastfeeding (OR = 0.39; 95% CI: 0.21-0.69) and continue breastfeeding until 3 months postpartum (OR = 0.35; 95% CI: 0.20-0.61) compared to food secure women, but the associations were no longer significant after adjustment for sociodemographic and health-related factors. Prenatal food insecurity was not associated with breastfeeding at 3 months postpartum in analyses stratified into high- and low-adversity groups. Conclusions: Prenatal food insecurity is a strong predictor of breastfeeding practices. Though not significantly associated with breastfeeding practices after adjustment, screening for prenatal food insecurity may help clinicians identify women who may need more supports to initiate and maintain breastfeeding. This thesis is dedicated to the families who participated in MARCH. Without your dedication, this work would not be possible. Thank you. iii ACKNOWLEDGEMENTS Completing this thesis would not have been possible without support from several insightful mentors. I am indebted to my advisor and committee chair, Dr. Jean Kerver, as she constantly supports my individual passions while also connecting me to new opportunities that broaden both my mind and skillset. Her insight and compassion are unmatched. I also owe thanks to Dr. Ana Vazquez for her statistical guidance, as she always goes the extra mile when answering my questions. I would also like to thank Dr. Katherine Alaimo for reminding me that my research should always be conducted through the lens of health equity. Lastly, I owe thanks to the Child Health Advances from Research with Mothers (CHARM) study office staff for their endless efforts to collect the data used for this thesis. iv TABLE OF CONTENTS LIST OF TABLES ....................................................................................................................................... vi LIST OF FIGURES .................................................................................................................................... vii KEY TO ABBREVIATIONS .................................................................................................................... viii CHAPTER 1: INTRODUCTION ................................................................................................................. 1 1.1 Social Determinants of Health ...................................................................................................... 1 1.1.1 Food Insecurity ............................................................................................................. 1 1.2 Infant Feeding Practices: The Importance of Breastfeeding......................................................... 3 1.3 Conceptual Framework ................................................................................................................ 4 CHAPTER 2: METHODS ............................................................................................................................ 6 2.1 Study Population .......................................................................................................................... 6 2.2 Data Collection ............................................................................................................................. 6 2.2.1 Food Insecurity ............................................................................................................. 6 2.2.2 Sociodemographic and Health-related Covariates ....................................................... 7 2.2.3 Breastfeeding Initiation and Duration ......................................................................... 8 2.3 Statistical Analysis ...................................................................................................................... 8 2.3.1 Multiple Logistic Regression ...................................................................................... 9 2.3.2 Cox-Proportional Hazards Ratio ................................................................................. 9 2.3.3 Adversity Index .......................................................................................................... 10 CHAPTER 3: RESULTS ........................................................................................................................... 12 3.1 Study Participants ....................................................................................................................... 12 3.2 Multiple Logistic Regression: Breastfeeding Initiation & Breastfeeding at 3 Months Postpartum .................................................................................................................................. 13 3.3 Cox-Proportional Hazards Ratio: Breastfeeding Duration ........................................................ 14 3.4 Adversity Index .......................................................................................................................... 14 CHAPTER 4: DISCUSSION ...................................................................................................................... 16 4.1 Principal Findings ...................................................................................................................... 16 4.2 Results in the Context of What is Known ................................................................................. 16 4.3 Strengths and Limitations ........................................................................................................... 18 4.4 Conclusions ................................................................................................................................ 19 APPENDICES ........................................................................................................................................... 21 APPENDIX A: Tables ...................................................................................................................... 22 APPENDIX B: Figures ..................................................................................................................... 33 REFERENCES ........................................................................................................................................... 38 v LIST OF TABLES Table 1. Sociodemographic and health-related characteristics in the analytic sample and by food insecurity status........................................................................................................................................... 23 Table 2. Sociodemographic and health-related characteristics in the analytic sample and by breastfeeding initiation status. ........................................................................................................................................... 25 Table 3. Sociodemographic and health-related characteristics in the analytic sample and by breastfeeding status at 89 days postpartum among those who initiated breastfeeding...................................................... 27 Table 4. Unadjusted and adjusted associations between prenatal food insecurity and breastfeeding outcomes. .................................................................................................................................................... 29 Table 5. Percent breastfeeding at 3 months postpartum by adversity score (N = 491). .............................. 30 Table 6. Descriptive characteristics of low-adversity (adversity score ≤ 2) and high-adversity (adversity score > 2) groups......................................................................................................................................... 31 Table 7. Associations between food insecurity and breastfeeding at 3 months postpartum within the high- adversity and low-adversity groups using Fisher’s Exact Test. .................................................................. 32 vi LIST OF FIGURES Figure 1. Summary of socio-ecological barriers to breastfeeding. ............................................................. 34 Figure 2. Derivation of the Analytic Sample. ............................................................................................. 35 Figure 3. Unadjusted Kaplan-Meier curve for breastfeeding duration in days by food insecurity status in the analytic sample (N = 495). Breastfeeding duration is censored at the infant’s age at the 3-month survey for those still breastfeeding at the time of the survey. ..................................................................... 36 Figure 4. Percent breastfeeding at 3 months postpartum by adversity score (N = 491). ............................ 37 vii KEY TO ABBREVIATIONS aHR Adjusted Hazard Ratio aOR Adjusted Odds Ratio BF Breastfed BMI Body Mass Index HP Health Plan HR Hazard Ratio MARCH Michigan Archive for Research on Child Health OR Odds Ratio WIC Special Supplemental Nutrition Program for Women, Infants, and Children viii CHAPTER 1: INTRODUCTION 1.1 Social Determinants of Health The social determinants of health, or the conditions where people live, learn, work, and play, are now recognized as a significant contributor to health inequities.1 Substantial evidence demonstrates that decreasing levels of income, education, social status, and social support are associated with increased illness and death throughout the lifespan.2,3 Such social determinants present in many ways, including discrimination, neighborhood safety, transportation, stress or allostatic load, limited access to quality care, housing insecurity, and food insecurity, and these factors often co-occur within individuals and communities. Some estimate that around 50% of one’s health status can be determined by these socio- economic factors.4,5 Despite growing evidence suggesting that social determinants are often root causes to medical problems, initiatives to address social determinants of health remain sparse, and are rarely integrated into standard medical care practices. A survey of primary care providers found that 85% believe that unmet social needs are leading directly to worse health outcomes among Americans, but 80% of physicians did not feel confident in their ability to meet their patients’ social needs.2,6 However, there is also a growing field of front-line public health workers who advocate for evidence-based guidance on how to better address social determinants of health, and many agree that screening for social determinants of health should now be integrated into primary care practice.2 1.1.1 Food Insecurity Food insecurity, defined as the lack of consistent access to enough food to sustain an active and healthy life, is just one social determinant of health.7 In 2020, Feeding America estimated that 1 in 8 Americans were food insecure, and households with children faced even higher prevalence of food insecurity.7,8 Though this high prevalence of food insecurity is alarming alone, it is of particular concern due to the myriad of physical and psychosocial health outcomes associated with food insecurity, including type 2 diabetes, obesity, and poor mental health. In fact, food insecurity is a stronger predictor of chronic disease than income.9 Though there is relatively limited research on the implications of food insecurity 1 during pregnancy, prenatal food insecurity has been shown to associate with decreased quality of life, poor psychosocial health, gestational weight gain, and pre-pregnancy body mass index (BMI).10–13 Food insecurity has a broad scope of impact at all stages of the life course, and given its high prevalence, more work addressing food insecurity is needed. Because families with children face higher rates of food insecurity, one would hypothesize that the peripartum period would be a critical window for the development of food insecurity and assessing and intervening on food insecurity prenatally may improve health outcomes for families with children. Despite this theoretical hypothesis, relatively little work has attempted to even measure prevalence of food insecurity during pregnancy in the general population.14 Data from the Pregnancy, Infection, and Nutrition cohort of women in North Carolina from 2001-2005 found that 14% of pregnant women were marginally food insecure and 10% were food insecure, but more nationally representative estimates of food insecurity specific to the prenatal period are not available.11 Though there is little recent research assessing the prevalence and impact of food insecurity in pregnancy, the American College of Obstetrics and Gynecology does suggest screening for food insecurity (among other social determinants of health) during pregnancy.15 Similar food insecurity screening recommendations have been provided by the American Academy of Pediatrics, who recommends universal food insecurity screening during pediatric care using a two-question screener called the Hunger Vital Sign. The two included questions are (1) “Within the past 12 months, we worried whether our food would run out before we got money to buy more” and (2) “Within the past 12 months, the food we bought just didn’t last and we didn’t have money to get more.” The goal of universal administration of the Hunger Vital Sign is to connect patients and their families to federal nutrition programs and food resources, document food insecurity status in medical records, and advocate for solutions to root causes of food insecurity all in an effort to promote child health.16 In addition to providing the Hunger Vital Sign screening resource, the American Academy of Pediatrics also provides extensive resources and education materials on how to implement food insecurity screening into 2 clinical care, which may also be helpful to prenatal care providers. Because prenatal visits are frequent and occur in a concentrated time, perinatal care may be an especially opportunistic time for screening.14,17 Both theory and previous research suggest that the perception of limited resources (e.g., food insecurity) impacts both decision-making and health-related behaviors, which may explain mechanisms linking food insecurity to poor health.18 Such potential mechanisms include reduced access to healthy foods, reliance on cyclic eating pattens (likely due to intermittent insufficient income), and severe stress causing metabolic disturbances.13,18 These altered health-related behaviors and decision making may also contribute to unhealthy feeding patterns as early as infancy.18 Research on WIC-eligible mothers of young children found that during times when the families were food insecure, the mothers increased their own restrained eating, and the mothers restrained eating associated with more restrictive and less responsive child feeding practices.19 Another study among low-income Hispanic mothers found that food insecure mothers were more likely to exhibit obesogenic restrictive and pressuring infant feeding styles.20 Because healthy infant feeding practices are integral to a child’s lifelong development, more research regarding the potential impact of food insecurity on a mother’s infant feeding practices is warranted. 1.2 Infant Feeding Practices: The Importance of Breastfeeding The American Academy of Pediatrics recommends exclusive breastfeeding (i.e., infant receives no other foods or formula besides breast milk) though infant age 6 months, as breastfeeding has been shown to benefit many realms of child health and development, including reduced risk of asthma, obesity, and infections.21,22 In the United States, however, only about 1 in 4 infants meet this recommendation, and large disparities in breastfeeding practices persist by sociodemographic group.22 According to the Centers for Disease Control and Prevention, fewer than 75.5% of non-Hispanic Black infants are ever breastfed (i.e., infant received any breastmilk), while 85.3% of non-Hispanic White infants are ever breastfed. Infants receiving Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) benefits are also less likely to be breastfed than those eligible but not receiving WIC (76.9% vs 81.2%). Despite efforts to increase breastfeeding rates, prevalence of any breastfeeding and exclusive breastfeeding through 6 months have only increased marginally in recent years.23 3 Causes of low adherence to breastfeeding recommendations and the persistence of breastfeeding disparities can begin to be understood when viewing breastfeeding through a socio-ecological lens. The socio-ecological model shifts the onus of health behavior decision making (e.g., decision to breastfeed) from the individual alone and includes intrapersonal, interpersonal, institutional, and community level factors that provide context for one’s decision.24 Qualitative work suggests that breastfeeding is a valued behavior and mothers desire to breastfeed, but multi-level barriers experienced by mothers often prevent them from being able to breastfeed as planned.25 These barriers are summarized in Figure 1. At the individual level, barriers include exhaustion, time commitment of breastfeeding, and feelings of isolation, while interpersonal factors include lack of family or partner support for breastfeeding. Community level barriers involve social support (e.g., lack of community support groups), but also include social acceptability of breastfeeding in public spaces. Organizational barriers often begin at the hospital (e.g., lack of certified lactation counselors, formula advertising, provided formula samples),25,26 and policy level factors include inadequate maternity leave and inability to pump breastmilk at work. 25 In order to improve breastfeeding disparities, policy and interventions must consider the wide array of barriers experienced by mothers. 1.3 Conceptual Framework Because of the multi-level social determinants of breastfeeding, and because of the impact food insecurity has on feeding practices more generally, we hypothesize that women with prenatal food insecurity may also endure more barriers to breastfeeding. Our hypothesis is consistent with qualitative findings, where low-income food insecure mothers reported actively limiting their portion sizes when eating and believed that their reliance on a poor diet combined with high stress levels would affect their breast milk quality.27 Observational research on the topic is limited. Most studies assessing associations between food insecurity and breastfeeding are cross-sectional and have mixed results, making it uncertain if food insecurity is indeed predictive of breastfeeding outcomes.20,28–36 Two cohort studies prospectively measured prenatal food insecurity, but were conducted in limited samples (low-income Hispanic mothers and WIC participants), warranting replication in more diverse populations.37,38 Previous studies also 4 model associations with different covariates, making results difficult to interpret, given the multiple socio-demographic and health-related barriers faced by food insecure mothers. To fill this gap in the literature, the present study assesses associations between prenatal food insecurity and breastfeeding behaviors in a diverse prospective cohort of pregnant women in Michigan. Because the prevalence of prenatal food insecurity is not well established, our first aim was to provide prenatal food insecurity prevalence estimates that may inform health policy in Michigan. Our second aim was to assess associations between prenatal food insecurity and (1) breastfeeding initiation, (2) percent breastfeeding at 3 months postpartum, and (3) breastfeeding duration until 3 months postpartum using sequential models that adjust for relevant covariates. Next, we created an adversity index to quantify the number of breastfeeding barriers mothers experienced to establish if the association between prenatal food insecurity and breastfeeding may differ in high-adversity facing and low-adversity facing groups. This approach is intended to describe prenatal food insecurity in the context of other factors disproportionately experienced by food-insecure mothers, which may help illustrate why screening for prenatal food insecurity may be beneficial. 5 CHAPTER 2: METHODS 2.1 Study Population The study population consisted of women who participated in the Michigan Archive for Research on Child Health (MARCH), an ongoing prospective population-based pregnancy cohort representing all births in Michigan’s lower peninsula. MARCH recruits women at their first prenatal visit at 22 Michigan clinics and includes follow-up of mother-infant dyads throughout early childhood. The goal of MARCH is to archive biospecimens and survey data for research on causes of adverse maternal and child outcomes. Eligibility criteria for MARCH include being age 18 years or older and being able to complete surveys in English. Recruitment began in 2017, and because recruitment is still ongoing, this study utilized data collected as of February 2022. Our analysis included mothers who provided informed consent, had singleton pregnancies, completed the food insecurity questions during the prenatal survey, and completed the three-month postpartum follow-up visit. This study was approved by the Michigan State University Institutional Review Board. 2.2 Data Collection 2.2.1 Food Insecurity Food insecurity was assessed prenatally using the following three questions: (1) During the past month, did you ever eat less than you felt you should because there wasn't enough money to buy food? (2) During the past month, did you ever get emergency food from a church, food pantry, food bank or eat in a food kitchen? (3) During the past month, have you been concerned about having enough food for you or your family? These questions are similar to questions asked on the Current Population Survey Food Security Supplement (CPS-FSS), but our measure uses a 30 day timeframe (as opposed to a 12 month timeframe) in order to measure food insecurity specific to the prenatal period.39 Mothers who answered yes to any of the three questions were categorized as food insecure, and mothers who said no to all three were categorized as food secure. Five participants who stated ‘don’t know’ were excluded from the analysis. 6 2.2.2 Sociodemographic and Health-related Covariates Several questionaries were administered to mothers via telephone during pregnancy through 3 months postpartum. Mother’s race, age, marital status, education, employment status, health plan type, household size, smoking status, and pregnancy intention were assessed during pregnancy around the mother’s first prenatal visit. For descriptive and analytic purposes, race was categorized as non-Hispanic Black, non-Hispanic White, and other race. Mother’s age at birth was categorized as ages 18-25 years, 26- 33 years, and greater than 33 years. Marital status was categorized as married, living with a partner, and single (divorced/separated/widowed/never married). Mother’s employment during the prenatal survey was categorized as working full time, part time, or not working for pay. Household size was assessed as the number of people relying on the household income, and was categorized as 1-2 people, 3-4 people, and 5 or more people. Health plan type was categorized as (1) having health insurance only through the government (e.g., Medicaid), (2) having health insurance from a job, spouse, or parent, and (3) having other, multiple, or no health insurance. Smoking was assessed around the first prenatal visit and categorized as non-smokers, those who quit since becoming pregnant, and those who still reported smoking. Pregnancy intention was assessed dichotomously as the response to “Was this pregnancy planned?”. Maternal Body Mass Index (BMI) was assessed from self-reported height and pre-pregnancy weight. BMI was categorized as underweight (< 18.5 kg/m2), healthy weight (18.5-24.99 kg/m2), overweight (25-29.99 kg/m2) and obese (≥30 kg/m2).40 Prenatal depression was also assessed at the prenatal survey using the validated Edinburgh Depression Scale, which was scored according the scale instructions.41 The Edinburgh Depression Scale has a potential score range of 0 to 30.41 A cut-off point of 11 or higher was used to indicate possible depression, as previous literature shows that this cut-off maximizes sensitivity and specificity in pregnant populations.42 Additional sociodemographic and health-related variables were collected at the 3-month postpartum study survey. The mother’s participation in the Women Infant and Children Supplemental Nutrition Program (WIC) was assessed as the mother’s self-reported receipt of WIC vouchers for herself or the 7 baby in the prior month. Hospital length of stay at birth (in days) and birth sex were also self-reported by the mother at the 3-month survey. Birth certificate data was also obtained for participants who consented to provide birth certificate data. Parity, birth sex, birth date, physician-estimated gestational age (in weeks), birth weight, and delivery route were abstracted from the birth certificates. Gestational age was dichotomized as those less than 37 weeks gestation to indicate preterm birth and those greater than or equal to 37 weeks gestation.43 Birth weight was dichotomized as those with birth weights less than 2500 grams (i.e., low birth weight) and those with birth weights greater than or equal to 2500 grams.44 For those in the analytic sample who did not provide birth certificate data (N = 35), gestational age was calculated (in weeks) as the time between the mother’s last menstrual period (reported at the prenatal survey) and the birth date. 2.2.3 Breastfeeding Initiation and Duration Information about breastfeeding practices were collected at the 3-month postpartum survey. The infant’s age at the survey was calculated in days, and the first survey occurred at 89 days postpartum. The mother was first asked the following question “Did (baby) ever have breast milk, including directly at the breast or from a bottle, or mixed in cereal or other foods?”. Mothers who said yes were categorized as breastfeeding initiators, and those who reported no were categorized as non-initiators. Mothers who initiated breastfeeding were then asked if they had completely stopped feeding the infant breastmilk both at the breast and via expressed milk in a bottle. Mothers who completely stopped breastfeeding reported the infant’s age in days, weeks, or months when they completely stopped breastfeeding and pumping milk. This age was converted to days by multiplying weeks by 7 and months by 30. Breastfeeding at 3 months postpartum was dichotomized as those still breasting at 89 days postpartum and those who were not breastfeeding at all at 89 days postpartum. 2.3 Statistical Analysis Chi-square tests and Fisher’s Exact tests (when expected number of cases in a cell was small) were utilized for covariate by exposure and outcome analyses. For all analyses, missing data was excluded in a 8 pairwise manner. All data analyses were performed using SAS software version 9.4 (SAS Institute, Cary NC). 2.3.1 Multiple Logistic Regression Consistent with methodology used in prior literature, unadjusted and multivariate logistic regressions were used to assess associations between food insecurity and the two primary outcomes: (1) breastfeeding initiation and (2) breastfeeding at 3 months postpartum among those who initiated breastfeeding. Six sequential models were analyzed for each outcome analysis to accommodate covariates. Model 1 was unadjusted. Models two through four added in the most relevant confounders that play a causal role in breastfeeding behaviors. Model two adjusted for pregnancy intention, as previous literature suggests a close relationship between pregnancy intentions and breastfeeding intentions.45,46 Model 3 added cigarette smoking, as mothers who smoke have lower breastfeeding rates; reasons for which have been documented as through feelings that it is unhealthy to breastfeed when smoking and through physiologic decreases in milk production.47,48 Model 4 added in marital status as a proxy for social support. Previous literature has shown that social support is associated with the mother’s ability to start and maintain breastfeeding. 49,50 Model 5 added in maternal education, as maternal education may play a direct role in the mother’s education about breastfeeding specifically.51 Model 6 added in mother race, age, and health plan type. These variables are less directly causally related to breastfeeding outcomes and may thus be proxies for other determinants of breastfeeding. Previous literature uses these variables, so model 6 serves to both replicate and ensure these results are comparable to prior literature.28,29,32,37,52 2.3.2 Cox-Proportional Hazards Ratio Cox-proportional hazard ratio methods were used to assess differences in time-to-breastfeeding cessation by food insecurity status in the entire analytic sample (i.e., breastfeeding initiators and never breast feeders). The same six iterative models were used as outlined in section 2.3.1. Log-log plots were used to assess proportionality of the hazards over time. Kaplan-Meier curves depicting time-to- breastfeeding cessation among all participants (breastfeeding initiators and never breast feeders) by food insecurity status were graphed. 9 2.3.3 Adversity Index Previous research has shown that the many different adversities influence a women’s ability to breastfeed, and these adversities most often do not occur in isolation.25 To illustrate if these adversities operate in an additive fashion, we followed the methods used by Alaimo et al. to create a risk factor index by summing the factors associated with not breastfeeding or early breastfeeding cessation.53 Participants with missing data in greater than one included category were excluded (N = 4). Each of the following factors were given one point and summed together to create an adversity score. These eight variables chosen for the adversity index were selected based on their availability in the data and the strength of the association between the factor and breastfeeding practices based on prior literature. 1. Age at birth less than 20 years: Previous research on adolescent parents has shown that parents less than age 20 endure lower breastfeeding rates compared to mothers aged 20 years or older (74% vs 82-84%). Many mechanisms have been proposed, including less social support, lack of school or work-based facilities to breastfeed, and lower breastfeeding education.26,54 2. Single marital status: Mothers with single marital status have also been shown to breastfeed at lower rates than married and co-habituating mothers, presumably through social support mechanisms.49,50 3. Maternal education of high school completion or less: Previous research suggests that there is a graded relationship between maternal education and breastfeeding initiation, with mothers with education of high-school or less having lower breastfeeding rates.51 4. WIC recipient: Though WIC provides additional incentives to women who choose to breastfeed, research consistently shows that WIC recipients experience lower ever-breastfeeding rates than WIC-eligible non-participants.55 5. Unplanned pregnancy: Pregnancy intention is also strongly associated with breastfeeding initiation and duration, potentially because decisions about breastfeeding intentions are made during pregnancy, and women with planned pregnancies often have stronger social support.45,46 10 6. BMI greater than or equal to 30 kg/m2: Women with obesity also have been shown to experience reduced breastfeeding rates. Proposed mechanisms include both physiologic factors (e.g., delayed lactogenesis, mechanical factors regarding additional body tissue, hormonal imbalance) and psychosocial factors (e.g., body image, depression, less control over breastfeeding due to having highly medicalized pregnancies).56–59 7. Government health insurance: Mothers on Medicaid (government) health plans also have been shown to experience lower odds of breastfeeding at 6-8 weeks postpartum than those with commercial insurance after accounting for other confounding factors.60 Barriers to care among Medicaid recipients have been reported as including limited access to facilities that accept Medicaid plans, limited availability of culturally competent care, and limited appointment times that accommodate schedules of low-income workers.61 8. Cigarette smoking during pregnancy: Additional research suggests that women who smoke during pregnancy also experience lower breastfeeding rates potentially due to both physiologic (e.g., reduced milk volume, shorter lactation period) and psychosocial (e.g., mixed messages regarding the healthfulness of breastmilk of smokers) mechanisms. 47,48 To demonstrate if these factors indeed operate in an additive fashion, the percent breastfeeding at 3 months postpartum was plotted against the adversity score. A chi-square test was used to assess the association between the adversity score and breastfeeding status at 3 months postpartum. Then the analytic sample was divided into a high-adversity (score > 2) and low-adversity group (score ≤ 2). The cut point of 2 was selected based on both visual inspection of the plot outlined above and based on selecting a score that split the analytic sample most equally to avoid small subgroups. To assess for interactions between food insecurity and these other adversities, Fisher’s Exact test was utilized to assess associations between food insecurity and breastfeeding at 3 months postpartum within the low-adversity and high- adversity groups. 11 CHAPTER 3: RESULTS 3.1 Study Participants As of February 5, 2022, 1,113 pregnant women consented to participate in MARCH, 892 of which participated in initial data collection (i.e., completed prenatal survey). After exclusions were applied (multiple-gestation pregnancies, missing food insecurity questions, 3-month survey not completed either because of a missing data point or because the participants are not yet in the study visit time window in this ongoing cohort study), the final analytic sample was 495 (Figure 2). Maternal and infant characteristics by food insecurity status are listed in Table 1. Sixteen percent of women reported prenatal food insecurity, and most women (86%) initiated breastfeeding. Of those who initiated breastfeeding, 63% were still breastfeeding at 3 months postpartum. Most participants were non-Hispanic White (61%), age 26-33 years, married (54%), had a bachelor’s degree of higher (43%) and worked full time during pregnancy (57%). Most women had health insurance through a job or family member (50%) and 42% were WIC recipients. Few participants reported smoking throughout pregnancy (11%) and slightly over half reported planned pregnancies (53%). Nineteen percent of women experienced prenatal symptoms consistent with possible depression, and over half of women were overweight or obese before pregnancy. Regarding birth outcomes, 90% of births occurred at 37 weeks or later, and 93% of infants had birth weights greater than or equal to 2500 grams. As is evident in Table 1, differences were found in most sociodemographic characteristics between food insecurity groups. A larger proportion of participants experiencing food insecurity were non-Hispanic black, ages 18-25 years, single, had education of high school or less, were not working for pay, and had a household size with 5 or more people. A larger proportion of participants experiencing food insecurity also had government health insurance, received WIC, smoked, had unplanned pregnancies, and were multiparous. A higher proportion of participants experiencing food insecurity had possible depression and obesity. No significant differences by food insecurity status were observed for birth outcomes (i.e., gestational age, birth weight, delivery route, hospital length of stay). 12 Table 2 displays sociodemographic and health-related characteristics by breastfeeding initiation status. A larger percentage of those who never breastfed were non-Hispanic black, ages 18-25 years, single, had education of high school or less, and had a household size with 5 or more people. A larger percentage of those who never breastfed also had government health insurance, received WIC, smoked, had obesity, and had unplanned pregnancies. There were no notable differences in parity, prenatal depression, or birth outcomes (i.e., gestational age, birth weight, delivery route, hospital length of stay) among those who initiated versus never breastfed. Table 3 shows sociodemographic and health-related characteristics by breastfeeding status at 3 months postpartum among those who initiated breastfeeding. A larger proportion of those who discontinued breastfeeding by 3 months postpartum were non-Hispanic black, ages 18-25 years, single, and had education of high school or less. A larger percentage of those who discontinued breastfeeding also had government health insurance, received WIC, smoked, had obesity, and had unplanned pregnancies. There were no notable differences in parity by breastfeeding status at 3 months postpartum, but a larger proportion of those who discontinued breastfeeding had possible depression (24% vs 15%). There were no differences in gestational age, birth weight, or delivery route between groups, but a larger proportion of those who discontinued breastfeeding by 3 months postpartum had hospital length of stays at delivery of 5 days or more (18% vs 9%). 3.2 Multiple Logistic Regression: Breastfeeding Initiation & Breastfeeding at 3 Months Postpartum In the unadjusted model, women who reported food insecurity during pregnancy were less likely to initiate breastfeeding compared to food secure women (OR = 0.39; 95% CI: 0.21-0.69). The association remained significant when pregnancy intention was added to the model (aOR = 0.49; 95% CI: 0.27-0.91) but was no longer significant in model 3 when cigarette smoking was added. Similar results were seen for the association between prenatal food insecurity and breastfeeding at 3 months postpartum among those who initiated breastfeeding. In the unadjusted model, women who reported prenatal food insecurity had lower odds of breastfeeding at 3 months postpartum (OR = 0.35; 95% CI: 0.20-0.61). The association remained significant when pregnancy intention was added to the model (aOR = 0.48; 95% CI: 13 0.27-0.87), but when cigarette smoking was added to the model, the association was no longer significant (Table 4). 3.3 Cox-Proportional Hazards Ratio: Breastfeeding Duration Using an unadjusted Cox-proportional hazards model of the analytic sample (breastfeeding initiators and never breast feeders), women who reported prenatal food insecurity experienced greater than twice the risk of breastfeeding cessation during the first three months postpartum when compared to food-secure women (HR = 2.29; 95% CI: 1.72-3.06). The association remained significant when pregnancy intention (aHR = 1.89; 95% CI: 1.33-2.43) and cigarette smoking (aHR = 1.62; 95% CI: 1.19- 2.20) were added to the model. The association was no longer significant when marital status was added to the model (Table 4). Figure 3 shows the unadjusted Kaplan-Meier curve for breastfeeding duration by food insecurity status. 3.4 Adversity Index Though the adversity index had a potential range from 0 to 8, the index score in our sample ranged from 0 to 7. Because only 9 participants had an index score equal to 7, scores of 6 and 7 were combined for descriptive and analytic purposes. Results from a chi-square test found the adversity score to be significantly associated with breastfeeding status at 3 months postpartum (p < 0.001). Twenty-eight percent of the sample had an adversity score of zero, and an additional 29% had an adversity score of 1-2 (Table 5). Those with an adversity score of 0 had the highest percentage breastfeeding at 3 months postpartum (78%) and the percentage breastfeeding at 3 months decreased with each additional adversity score point (Figure 4). Those with an adversity score of 6 or more had the lowest percentage breastfeeding at 3 months postpartum (16%). The sample was dichotomized into a low-adversity (adversity score ≤ 2; N = 283) versus high- adversity (adversity score > 2; N = 208) groups. A larger proportion of the high-adversity group was food insecure (30.2% vs. 5.3%; p < 0.0001), non-Hispanic Black (59.1% vs. 6.7%; p < 0.0001) and were not breastfeeding at 3 months postpartum (70.7% vs. 27.2%; p < 0.0001) (Table 6). Fisher’s Exact test 14 demonstrated that prenatal food insecurity was not associated with breastfeeding at age 3 months in either the low-adversity group or the high-adversity group (Table 7). 15 CHAPTER 4: DISCUSSION 4.1 Principal Findings In this diverse sample of Michigan pregnant women, nearly one in six women experienced food insecurity during pregnancy, which is higher than the national average for U.S. adults (1 in 9) and comparable to average for children (1 in 7).7,8 Breastfeeding initiation rates (86%) in our sample were also comparable to the national average (84%).22 Prenatal food insecurity was strongly associated with both breastfeeding initiation and breastfeeding duration until 3 months postpartum in unadjusted analyses, but results were attenuated when important covariates were considered. Though these results suggest that prenatal food insecurity alone is not causally related to breastfeeding behaviors, the adversity index analysis demonstrated that a larger proportion of women facing many other barriers to breastfeeding endure food insecurity when compared to mothers with few breastfeeding barriers. Thus, food insecurity does not occur independently from other barriers, and screening for prenatal food insecurity may still be a useful tool in predicting the totality of the barriers endured by women and the impact that those multiple barriers may have on breastfeeding behaviors. 4.2 Results in the Context of What is Known Previous literature assessing associations between food insecurity and breastfeeding behaviors has been mixed. Most prior work has been cross-sectional during infancy, and most studies found significant associations between food insecurity and breastfeeding initiation before adjustment, but not after adjustment, which is consistent with our findings.28–30,33 However, prior studies included many covariates in adjusted models without theoretical explanations for their modeling choices, and we believe that some covariates used may be only proxies of potential confounders (e.g., race and income may be proxies for groups who experience more exhaustion and time constraints) and other covariates used may be mediators between food insecurity and breastfeeding (e.g., maternal depression, maternal BMI). Because of this possibility for overadjustment, our sequential addition of covariates and theoretical explanations for our model building strategy provides more information on which key covariates may be explaining the association. 16 Both food insecurity and breastfeeding behaviors are culturally specific, thus associations between them may operate differently in different populations. For example, studies in the Canadian Intuit population found that food insecurity was not associated with breastfeeding initiation or exclusivity even before adjusting for any covariates.33,35 Similarly, a study of low-income Hispanic mothers who were WIC participants found no association between food insecurity during pregnancy or infancy and breastfeeding duration, exclusivity, or breastfeeding status at 9 months postpartum.20,38 A U.S. cross- sectional study of primarily low-income minority families also found no differences in percent breastfeeding at 2 months postpartum between food secure and food insecure mothers in unadjusted analyses,36 further demonstrating the inconsistent and possibly sample-specific results of prior studies. Few previous studies found significant associations between food insecurity and breastfeeding behaviors after adjustments. The Canadian community survey found that those with household food insecurity had no differences in breastfeeding initiation compared to food secure households but did have significantly lower odds of exclusive breastfeeding at age 4 months after adjustment.31 They also found that those mothers with severe food insecurity breastfed for a significantly shorter time period (1.2 months, p = 0.04) than food secure mothers, but there was no difference for mothers with marginal or moderate food security.31 A small study of low-income patients at two Medicaid pediatric clinics found that food insecure mothers had a lower likelihood of breastfeeding at 2 months postpartum after adjustments.32 Another study of WIC participants found that low prenatal food security was associated with lower likelihood of initiating breastfeeding, but not breastfeeding duration.37 These results differed from our present study likely due to having more specific food insecurity measures, studying different populations, and having breastfeeding measures (e.g., exclusivity) that were not assessed in our analysis. Our study builds upon these prior studies by attempting to view food insecurity in the context of other adversities, instead of merely adjusting for these factors. Moreover, the adversity index creates a summation score that represents the totality of breastfeeding barriers endured by mothers, which helps illustrate breastfeeding differences in accord with the socio-ecological model. Our results suggest that mothers who face larger number of adversities have lower breastfeeding rates at 3 months postpartum. 17 This finding alludes to the difficulty we as researchers face when trying to parse out the effect of one social determinant of health from others, because these factors often co-occur. Furthermore, this leads us to question the practicality of adjusting for many covariates, as doing so may create a result that will not apply to real life scenarios and may mis-inform public health policy. Previous research in pediatric populations demonstrates that screening for food insecurity, using measures like the Hunger Vital Sign, during clinical care is feasible and effective at connecting families to resources. The Hunger Vital Sign has been shown to have high sensitivity (96.7%) and specificity (86.2%) when administered to low-income urban families with children in emergency departments and primary care settings.62 A cluster randomized controlled trial of mothers with infants at eight urban community health centers found that an intervention that involved screening for social determinants of health (including food insecurity) and providing information on community resources was associated with increased enrollment in community resources in the intervention group compared to the usual care group, suggesting the potential effectiveness of screening.63 Some preliminary research using a convenience sample of adults at emergency departments found that the Hunger Vital Sign was also highly sensitive (94%) in adult populations.64 Though minimal research has used the Hunger Vital Sign in prenatal populations, one study among pregnant women taking opioid agonist treatment for opioid use disorder found that prenatal food insecurity (measured using the Hunger Vital Sign) associated with increased risk of severe neonatal abstinence syndrome.65 Given that the Hunger Vital Sign is shorter and easier to administer compared to longer validated food insecurity measures, implementing similar screening policies into prenatal care may be feasible and effective at reducing poor health outcomes by connecting women and families to resources. 4.3 Strengths and Limitations There are several strengths to this study. First, this study includes a diverse sample recruited from 22 prenatal clinics in Michigan, enhancing the generalizability of the results. The prospective nature of the study is also a considerable strength, given that only two prior studies assessed prenatal food insecurity prospectively. Prospective collection of food insecurity may be more accurate than mothers’ 18 retrospective recall of prenatal food insecurity during infancy. Additionally, our food insecurity questions measured food insecurity during the prior 30 days, which gives a very acute picture of prenatal food insecurity compared to other studies which may assess food insecurity over 12 months. Lastly, our detailed descriptions of our covariate selection techniques and the inclusion of the adversity index provide more context to the association between prenatal food insecurity and breastfeeding behaviors. Our study also has several limitations. First, we did not have a measure of food insecurity during infancy. Because mother’s food insecurity status may change after acquiring WIC benefits, it would have been beneficial to describe relationships between both prenatal and postnatal food insecurity and breastfeeding behaviors. Additionally, the food insecurity measure used was not able to differentiate between very low, low, and marginal levels of food insecurity, which may have differing impacts on breastfeeding. Second, our analysis only includes breastfeeding duration data until 3 months postpartum, and it does not account for breastfeeding exclusivity or intensity. Third, we did not have data on several factors that would have been important to consider, including other material hardships (e.g., housing insecurity, difficulty paying bills), mothers’ access to maternity leave, and the mothers’ employment status after birth. Additionally, the adversity index weights all included factors equally, which may not accurately portray the impact each factor has on breastfeeding. Lastly, while our sample size was moderate, this study should be replicated with larger sample sizes. 4.4 Conclusions These results should inform future intervention studies that seek to improve breastfeeding outcomes by addressing social determinants. As is shown in our results, breastfeeding barriers that exist on multiple socio-ecological levels often co-exist within individuals and addressing any one barrier may not make a substantial impact on the outcome without considering the full context of barriers faced. This may also explain why programs like WIC, which seeks to enhance mother and child outcomes with food insecurity by providing food vouchers, have been shown to be ineffective at increasing breastfeeding rates.66 Moreover, WIC is designed to primarily address food insecurity, without addressing the various other stressors that women with food insecurity face. 19 Because prenatal food insecurity is predictive of both breastfeeding initiation and duration, screening for food insecurity during prenatal care may be an effective strategy for identifying women who likely face multiple barriers to breastfeeding. In our study, our three-item food insecurity questionnaire was well-accepted by participants, as only 5 women had to be excluded due to selecting “don’t know” as an answer. This shows that most women are aware of their food insecurity status and feel comfortable providing that information, which makes food insecurity a potentially better screener than some other socioeconomic indicators that women may feel less comfortable sharing (e.g., income). By using a simple food insecurity screener, like the two-question Hunger Vital Sign screener recommended by the American Academy of Pediatrics, clinicians trained in social determinants of health will be able to connect food insecure women with supports that may better enable mothers to meet their breastfeeding goals. 20 APPENDICES 21 APPENDIX A Tables 22 Table 1. Sociodemographic and health-related characteristics in the analytic sample and by food insecurity status. All Food Food (N = 495) Secure Insecure (N = 416) (N = 79) Category N (%) N (%) N (%) P Race/ethnicity Non-Hispanic Black 142 (29) 98 (24) 44 (56) <.0001 Non-Hispanic White 299 (61) 273 (66) 26 (33) Other 52 (11) 44 (11) 8 (10) Mother Age (years) 18-<26 118 (24) 86 (21) 32 (41) 0.0006 26-<34 251 (51) 222 (53) 29 (37) ≥34 126 (25) 108 (26) 18 (23) Marital Status Married 264 (54) 250 (60) 14 (18) <.0001 Living with Partner 101 (20) 82 (20) 19 (24) Divorced, Separated, Widowed, Never Married 128 (26) 83 (20) 45 (58) Education < High School 41 (8) 22 (5) 19 (25) <.0001 High school graduate, diploma, or GED 90 (18) 65 (16) 25 (32) Some college/technical/Associate's 147 (30) 119 (29) 28 (36) Bachelor's or Higher 213 (43) 208 (50) 5 (6) Mother Employment Full time 279 (57) 254 (61) 25 (32) <.0001 Part time 88 (18) 68 (16) 20 (26) Not working for pay 126 (26) 93 (22) 33 (42) Household Size (Number of people) 1-2 210 (45) 179 (45) 31 (46) 0.0159 3-4 199 (43) 177 (44) 22 (32) 5 or more 57 (12) 42 (11) 15 (22) Health plan (HP) Type HP from Job, Spouse, or Parent 248 (50) 240 (58) 8 (10) <.0001 Government 197 (40) 137 (33) 60 (77) Other, Multiple, None 48 (10) 38 (9) 10 (13) WIC Status Yes 209 (42) 144 (35) 65 (82) <.0001 No 283 (58) 269 (65) 14 (18) Smoking status Non-Smoker 384 (78) 342 (83) 42 (54) <.0001 Quit since becoming pregnant 55 (11) 38 (9) 17 (22) Smoker 53 (11) 34 (8) 19 (24) 23 Table 1 (cont’d) All Food Food (N = 495) Secure Insecure (N = 416) (N = 79) Category N (%) N (%) N (%) P Planned Pregnancy Yes 255 (53) 233 (57) 22 (28) <.0001 No 229 (47) 173 (43) 56 (72) Parity Nulliparous 141 (31) 122 (32) 19 (25) 0.3689 Primiparous 133 (29) 111 (29) 22 (29) Multiparous 186 (40) 150 (39) 36 (47) Prenatal Depression Depression not likely 395 (81) 353 (86) 42 (55) <.0001 Possible Depression 94 (19) 59 (14) 35 (45) 2 Body Mass Index (BMI) kg/m Underweight (<18.5) 19 (4) 14 (3) 5 (7) 0.0248 Healthy (18.5-<25) 188 (39) 168 (41) 20 (26) Overweight (25-<30) 110 (23) 95 (23) 15 (20) Obese (≥30) 171 (35) 135 (33) 36 (47) Birth Sex Male 233 (47) 199 (48) 34 (43) 0.4334 Female 262 (53) 217 (52) 45 (57) Gestational Age (Weeks) <37 weeks 49 (10) 42 (10) 7 (9) 0.7372 ≥37 weeks 440 (90) 369 (90) 71 (91) Birth weight (g) < 2500 g 30 (7) 24 (6) 6 (8) 0.6207 ≥2500 g 430 (93) 359 (94) 71 (92) Delivery Route Cesarean 155 (34) 128 (33) 27 (35) 0.7806 Vaginal 305 (66) 255 (67) 50 (65) Hospital Length of Stay at Delivery (days) 1-2 280 (57) 240 (58) 40 (51) 0.1383 3-4 147 (30) 124 (30) 23 (29) ≥5 66 (13) 50 (12) 16 (20) Missing data: Race/ethnicity (N = 2); Marital Status (N = 2); Education (N = 4); Mother employment (N = 2); Household Size (N = 29); Health Plan Type (N = 2); WIC Status (N = 3); Smoking Status (N = 3); Planned Pregnancy (N = 11; Parity (N = 35); BMI (N = 7); Gestational age (N = 6); Birth weight (N = 35); Delivery Route (N = 35); Hospital Length of Stay at Delivery (N = 2) 24 Table 2. Sociodemographic and health-related characteristics in the analytic sample and by breastfeeding initiation status. All Ever BF Never (N = 495) (N = 427) BF (N = 68) Category N (%) N (%) N (%) P Race/ethnicity Non-Hispanic Black 142 (29) 100 (23) 42 (63) <.0001 Non-Hispanic White 299 (61) 280 (66) 19 (28) Other 52 (11) 46 (11) 6 (9) Mother Age (years) 18-<26 118 (24) 94 (22) 24 (35) 0.019 26-<34 251 (51) 217 (51) 34 (50) ≥34 126 (25) 116 (27) 10 (15) Marital Status Married 264 (54) 256 (60) 8 (12) <.0001 Living with Partner 101 (20) 84 (20) 17 (25) Divorced, Separated, Widowed, Never Married 128 (26) 86 (20) 42 (63) Education < High School 41 (8) 25 (6) 16 (24) <.0001 High school graduate, diploma, or GED 90 (18) 63 (15) 27 (41) Some college/technical/Associate's 147 (30) 126 (30) 21 (32) Bachelor's or Higher 213 (43) 211 (50) 2 (3) Mother Employment Full time 279 (57) 250 (59) 29 (43) 0.0604 Part time 88 (18) 72 (17) 16 (24) Not working for pay 126 (26) 104 (24) 22 (33) Household Size (Number of people) 1-2 210 (45) 181 (45) 29 (46) 0.0573 3-4 199 (43) 178 (44) 21 (33) 5 or more 57 (12) 44 (11) 13 (21) Health plan (HP) Type HP from Job, Spouse, or Parent 248 (50) 238 (56) 10 (15) <.0001 Government 197 (40) 144 (34) 53 (79) Other, Multiple, None 48 (10) 44 (10) 4 (6) WIC Status Yes 209 (42) 155 (36) 54 (82) <.0001 No 283 (58) 271 (64) 12 (18) Smoking status Non-Smoker 384 (78) 344 (81) 40 (60) 0.0003 Quit since becoming pregnant 55 (11) 43 (10) 12 (18) Smoker 53 (11) 38 (9) 15 (22) 25 Table 2 (cont’d) All Ever BF Never (N = 495) (N = 427) BF (N = 68) Category N (%) N (%) N (%) P Planned Pregnancy Yes 255 (53) 236 (57) 19 (28) <.0001 No 229 (47) 181 (43) 48 (72) Parity Nulliparous 141 (31) 121 (30) 20 (32) 0.6138 Primiparous 133 (29) 118 (30) 15 (24) Multiparous 186 (40) 158 (40) 28 (44) Prenatal Depression Depression not likely 395 (81) 346 (82) 49 (74) 0.1475 Possible Depression 94 (19) 77 (18) 17 (26) 2 Body Mass Index (BMI) kg/m Underweight (<18.5) 19 (4) 18 (4) 1 (2) 0.0084 Healthy (18.5-<25) 188 (39) 167 (40) 21 (32) Overweight (25-<30) 110 (23) 101 (24) 9 (14) Obese (≥30) 171 (35) 136 (32) 35 (53) Birth Sex Male 233 (47) 208 (49) 25 (37) 0.0668 Female 262 (53) 219 (51) 43 (63) Gestational Age (Weeks) <37 weeks 49 (10) 42 (10) 7 (10) 0.9002 ≥37 weeks 440 (90) 380 (90) 60 (90) Birth weight (g) < 2500 g 30 (7) 25 (6) 5 (8) 0.6244 ≥2500 g 430 (93) 372 (94) 58 (92) Delivery Route Cesarean 155 (34) 129 (32) 26 (41) 0.171 Vaginal 305 (66) 268 (68) 37 (59) Hospital Length of Stay at Delivery (days) 1-2 280 (57) 248 (58) 32 (48) 0.1925 3-4 147 (30) 126 (30) 21 (32) ≥5 66 (13) 53 (12) 13 (20) Missing data: Race/ethnicity (N = 2); Marital Status (N = 2); Education (N = 4); Mother employment (N = 2); Household Size (N = 29); Health Plan Type (N = 2); WIC Status (N = 3); Smoking Status (N = 3); Planned Pregnancy (N = 11; Parity (N = 35); BMI (N = 7); Gestational age (N = 6); Birth weight (N = 35); Delivery Route (N = 35); Hospital Length of Stay at Delivery (N = 2) Abbreviations: BF = Breastfed 26 Table 3. Sociodemographic and health-related characteristics in the analytic sample and by breastfeeding status at 89 days postpartum among those who initiated breastfeeding. All Still BF at 89 Discont’ BF (N = 495) days before 89 days (N = 268) (N = 159) Category N (%) N (%) N (%) P Race/ethnicity Non-Hispanic Black 142 (29) 42 (16) 58 (37) <.0001 Non-Hispanic White 299 (61) 195 (73) 85 (54) Other 52 (11) 31 (12) 15 (9) Mother Age (years) 18-<26 118 (24) 43 (16) 51 (32) 0.0005 26-<34 251 (51) 144 (54) 73 (46) ≥34 126 (25) 81 (30) 35 (22) Marital Status Married 264 (54) 194 (72) 62 (39) <.0001 Living with Partner 101 (20) 45 (17) 39 (25) Divorced, Separated, Widowed, Never Married 128 (26) 29 (11) 57 (36) Education < High School 41 (8) 9 (3) 16 (10) <.0001 High school graduate, diploma, or GED 90 (18) 30 (11) 33 (21) Some college/technical/Associate's 147 (30) 71 (26) 55 (35) Bachelor's or Higher 213 (43) 158 (59) 53 (34) Mother Employment Full time 279 (57) 166 (62) 84 (53) 0.0144 Part time 88 (18) 49 (18) 23 (15) Not working for pay 126 (26) 53 (20) 51 (32) Household Size (Number of people) 1-2 210 (45) 109 (42) 72 (49) 0.2064 3-4 199 (43) 122 (47) 56 (38) 5 or more 57 (12) 26 (10) 18 (12) Health plan (HP) Type HP from Job, Spouse, or Parent 248 (50) 171 (64) 67 (42) <.0001 Government 197 (40) 73 (27) 71 (45) Other, Multiple, None 48 (10) 24 (9) 20 (13) WIC Status Yes 209 (42) 69 (26) 86 (54) <.0001 No 283 (58) 198 (74) 73 (46) Smoking status Non-Smoker 384 (78) 232 (87) 112 (71) <.0001 Quit since becoming pregnant 55 (11) 24 (9) 19 (12) Smoker 53 (11) 11 (4) 27 (17) 27 Table 3 (cont’d) All Still BF at 89 Discont’ BF (N = 495) days before 89 days (N = 268) (N = 159) Category N (%) N (%) N (%) P Planned Pregnancy Yes 255 (53) 175 (67) 61 (39) <.0001 No 229 (47) 87 (33) 94 (61) Parity Nulliparous 141 (31) 75 (30) 46 (32) 0.6028 Primiparous 133 (29) 79 (31) 39 (27) Multiparous 186 (40) 97 (39) 61 (42) Prenatal Depression Depression not likely 395 (81) 226 (85) 120 (76) 0.0281 Possible Depression 94 (19) 40 (15) 37 (24) 2 Body Mass Index (BMI) kg/m Underweight (<18.5) 19 (4) 13 (5) 5 (3) 0.004 Healthy (18.5-<25) 188 (39) 120 (45) 47 (30) Overweight (25-<30) 110 (23) 61 (23) 40 (25) Obese (≥30) 171 (35) 71 (27) 65 (41) Birth Sex Male 233 (47) 129 (48) 79 (50) 0.7565 Female 262 (53) 139 (52) 80 (50) Gestational Age (Weeks) <37 weeks 49 (10) 23 (9) 19 (12) 0.2712 ≥37 weeks 440 (90) 241 (91) 139 (88) Birth weight (g) < 2500 g 30 (7) 13 (5) 12 (8) 0.2292 ≥2500 g 430 (93) 238 (95) 134 (92) Delivery Route Cesarean 155 (34) 75 (30) 54 (37) 0.1449 Vaginal 305 (66) 176 (70) 92 (63) Hospital Length of Stay at Delivery (days) 1-2 280 (57) 176 (66) 72 (45) 0.0001 3-4 147 (30) 67 (25) 59 (37) ≥5 66 (13) 25 (9) 28 (18) Missing data: Race/ethnicity (N = 2); Marital Status (N = 2); Education (N = 4); Mother employment (N = 2); Household Size (N = 29); Health Plan Type (N = 2); WIC Status (N = 3); Smoking Status (N = 3); Planned Pregnancy (N = 11; Parity (N = 35); BMI (N = 7); Gestational age (N = 6); Birth weight (N = 35); Delivery Route (N = 35); Hospital Length of Stay at Delivery (N = 2) Abbreviations: BF = Breastfed 28 Table 4. Unadjusted and adjusted associations between prenatal food insecurity and breastfeeding outcomes. Breastfeeding Still Breastfeeding Initiationa Breastfeeding at 3 Durationc b (Yes vs No) Months (Yes vs No) OR (95% CI) OR (95% CI) HR (95%CI) Model 1: Food Insecurity Status 0.39 (0.21, 0.69)* 0.35 (0.20, 0.61)* 2.29 (1.72, 3.06)* (Unadjusted)d Model 2: Model 1 + Pregnancy 0.49 (0.27, 0.91)* 0.48 (0.27, 0.87)* 1.80 (1.33, 2.43)* Intention Model 3: Model 2 + Cigarette 0.58 (0.31, 1.09) 0.56 (0.30, 1.02) 1.62 (1.19, 2.20)* Smoking Model 4: Model 3 + Marital Status 0.88 (0.46, 1.67) 0.69 (0.36, 1.31) 1.31 (0.96, 1.78) Model 5: Model 4 + Maternal 1.09 (0.56, 2.13) 0.74 (0.38, 1.42) 1.19 (0.87, 1.64) Education Model 6: Model 5 + Maternal Race 1.20 (0.60, 2.50) 0.72 (0.37, 1.41) 1.18 (0.85, 1.62) + Maternal Age + Health Plan Type a Logistic Regression of breastfeeding initiation among all participants (N = 495) b Logistic regression of breastfeeding at 3 months postpartum among those who ever initiated breastfeeding (N = 427) c Cox proportional hazards model for the hazard of stopping breastfeeding from birth until 3 months postpartum among all participants (N = 495) d Food Insecurity is measured dichotomous as food secure vs food insecure. Referent = Food Secure * p < 0.05 29 Table 5. Percent breastfeeding at 3 months postpartum by adversity score (N = 491).* Adversity All Participants Still breastfeeding at 3 months postpartum among the Score analytic sample N (%) N (%) Pa 0 139 (28) 108 (78) <.0001 1 83 (17) 57 (69) 2 61 (12) 41 (67) 3 49 (10) 22 (45) 4 66 (13) 22 (33) 5 50 (10) 10 (20) ≥6 43 (9) 7 (16) * The analytic sample consisted of N = 495, but N = 4 were deleted from the adversity score analysis due to missing data on greater than 1 factor included in the score composite. Each of the following factors were given a point of one each and summed together to create an adversity score: age at birth less than 20, single marital status, maternal education of high school completion or less, WIC recipient, unplanned pregnancy, BMI greater than or equal to 30 kg/m2, government health insurance, and cigarette smoking during pregnancy. a Chi-square test 30 Table 6. Descriptive characteristics of low-adversity (adversity score ≤ 2) and high-adversity (adversity score > 2) groups. Low- High-Adversity Adversity Group Group (N = 283) (N = 208) N (%) N (%) Pa Food Insecurity Food Insecure 15 (5.3) 63 (30.2) <.0001 Food Secure 268 (95.6) 145 (69.7) Still BF at 3 Months No 77 (27.2) 147 (70.7) <.0001 Yes 206 (72.8) 61 (29.3) Race Non-Hispanic Black 19 (6.7) 123 (59.1) <.0001 Non-Hispanic White 234 (82.7) 65 (31.3) Other 30 (10.6) 20 (9.6) a Chi-square test 31 Table 7. Associations between food insecurity and breastfeeding at 3 months postpartum within the high- adversity and low-adversity groups using Fisher’s Exact Test. Still Breastfeeding at 3 Months Postpartum Low-Adversity Group High-Adversity Group (N = 283) (N = 208) Still BF: Still BF: Still BF: Still BF: Yes No Yes No N N P N N P Food Insecurity Food Insecure 9 6 0.2473 15 48 0.3201 Food Secure 197 71 46 99 32 APPENDIX B Figures 33 Figure 1. Summary of socio-ecological barriers to breastfeeding.* Policy Barriers •Inadequate maternity leave •Limited ability to pump breastmilk at work •Hospital resources (e.g., lack of certified Organizational Barriers lactation counselors, formula advertising, provided formula samples) •Lack of normalization of breastfeeding in public Community Barriers •Lack of resource access (e.g., community support groups) Interpersonal Barriers •Lack of partner or family support •Isolation Individual Barriers •Exhaustion •Time commitment of breastfeeding * This figure summarizes results from Bookhart et al. (2021), Snyder et al. (2021), and the American College of Obstetrics and Gynecology Breastfeeding Expert Work Group (2021).24–26 34 Figure 2. Derivation of the Analytic Sample Pregnant Women Who Exclusion Criteria Initially Consented to MARCH N = 1113 Multiple-Gestation Pregnancy: N = 21 Singleton Pregnancy N = 1092 Did not complete prenatal survey: N = 221 Responded Don’t Know or Refused Food Insecurity Questions: N = 5 Completed Prenatal Food Insecurity Measure Infant Not Yet Born: N = 174 N = 866 Not yet eligible for 3 Month Survey or still in window for completion but not completed: N = 35 Completed 3 Month Survey Refused Survey or Unable to Contact: N = 495 N = 162 *MARCH data as of 02/05/2022 35 Figure 3: Unadjusted Kaplan-Meier curve for breastfeeding duration in days by food insecurity status in the analytic sample (N = 495). Breastfeeding duration is censored at the infant’s age at the 3-month survey for those still breastfeeding at the time of the survey. 36 Figure 4. Percent breastfeeding at 3 months postpartum by adversity score (N = 491).* * The analytic sample consisted of N = 495, but N = 4 were deleted from the adversity score analysis due to missing data on greater than 1 factor included in the score composite. Each of the following factors were given a point of one each and summed together to create an adversity score: age at birth less than 20, single marital status, maternal education of high school completion or less, WIC recipient, unplanned pregnancy, BMI greater than or equal to 30 kg/m2, government health insurance, and cigarette smoking during pregnancy. 37 REFERENCES 38 REFERENCES 1. About Social Determinants of Health (SDOH). Centers for Disease Control and Prevention. Published March 10, 2021. Accessed April 2, 2022. https://www.cdc.gov/socialdeterminants/about.html 2. Andermann A. Screening for social determinants of health in clinical care: moving from the margins to the mainstream. Public Health Rev. 2018;39:19. doi:10.1186/s40985-018-0094-7 3. CSDH. Closing the Gap in a Generation: Health Equity through Action on the Social Determinants of Health. Final Report of the Commission on Social Determinants of Health. World Health Organization; 2010:253. Accessed April 3, 2022. http://www.revistas.usp.br/rdisan/article/view/13190 4. Marmot M, Allen JJ. Social Determinants of Health Equity. Am J Public Health. 2014;104(Suppl 4):S517-S519. doi:10.2105/AJPH.2014.302200 5. Broader determinants of health. The King’s Fund. Accessed April 3, 2022. https://www.kingsfund.org.uk/projects/time-think-differently/trends-broader-determinants-health 6. Health Care’s Blind Side. RWJF. Published December 1, 2011. Accessed April 3, 2022. https://www.rwjf.org/en/library/research/2011/12/health-care-s-blind-side.html 7. Feeding America. What Is Food Insecurity in America? Hunger and Health: Feeding America. Accessed April 3, 2022. https://hungerandhealth.feedingamerica.org/understand-food-insecurity/ 8. Feeding America. Food Insecurity Report Briefs. Feeding America. Accessed April 3, 2022. https://www.feedingamerica.org/research/map-the-meal-gap/overall-executive-summary 9. Gregory CA. Food Insecurity, Chronic Disease, and Health Among Working-Age Adults. United States Department of Agriculture; 2017:31. 10. Moafi F, Kazemi F, Samiei Siboni F, Alimoradi Z. The relationship between food security and quality of life among pregnant women. BMC Pregnancy and Childbirth. 2018;18(1):319. doi:10.1186/s12884-018-1947-2 11. Laraia BA, Siega-Riz AM, Gundersen C, Dole N. Psychosocial Factors and Socioeconomic Indicators Are Associated with Household Food Insecurity among Pregnant Women. The Journal of Nutrition. 2006;136(1):177-182. doi:10.1093/jn/136.1.177 12. Di Renzo GC, Tosto V. Food insecurity, food deserts, reproduction and pregnancy: we should alert from now. J Matern Fetal Neonatal Med. Published online December 17, 2021:1-3. doi:10.1080/14767058.2021.2016052 13. Laraia BA, Siega-Riz AM, Gundersen C. Household Food Insecurity Is Associated with Self- Reported Pregravid Weight Status, Gestational Weight Gain, and Pregnancy Complications. Journal of the American Dietetic Association. 2010;110(5):692-701. doi:10.1016/j.jada.2010.02.014 39 14. Dolin CD, Compher CC, Oh JK, Durnwald CP. Pregnant and hungry: addressing food insecurity in pregnant women during the COVID-19 pandemic in the United States. American Journal of Obstetrics & Gynecology MFM. 2021;3(4). doi:10.1016/j.ajogmf.2021.100378 15. Committee on Health Care for Underserved Women. ACOG Committee Opinion No. 729: Importance of Social Determinants of Health and Cultural Awareness in the Delivery of Reproductive Health Care. Published 2018. Accessed April 19, 2022. https://journals.lww.com/greenjournal/Fulltext/2018/01000/ACOG_Committee_Opinion_No__729 __Importance_of.42.aspx 16. Screen and Intervene: A Toolkit for Pediatricians to Address Food Insecurity. Food Research & Action Center. Accessed April 27, 2022. https://frac.org/aaptoolkit 17. Reyes AM, Akanyirige PW, Wishart D, et al. Interventions Addressing Social Needs in Perinatal Care: A Systematic Review. Health Equity. 2021;5(1):100-118. doi:10.1089/heq.2020.0051 18. Laraia BA. Food Insecurity and Chronic Disease. Advances in Nutrition. 2013;4(2):203-212. doi:10.3945/an.112.003277 19. Armstrong B, Hepworth AD, Black MM. Hunger in the household: Food insecurity and associations with maternal eating and toddler feeding. Pediatr Obes. 2020;15(10):e12637. doi:10.1111/ijpo.12637 20. Gross RS, Mendelsohn AL, Fierman AH, Racine AD, Messito MJ. Food insecurity and obesogenic maternal infant feeding styles and practices in low-income families. Pediatrics. 2012;130(2):254- 261. doi:10.1542/peds.2011-3588 21. CDC. Facts About Nationwide Breastfeeding Goals. Centers for Disease Control and Prevention. Published August 24, 2021. Accessed April 3, 2022. https://www.cdc.gov/breastfeeding/data/facts.html 22. CDC. Why It Matters. Centers for Disease Control and Prevention. Published August 23, 2021. Accessed December 14, 2021. https://www.cdc.gov/breastfeeding/about-breastfeeding/why-it- matters.html 23. CDC. Results: Breastfeeding Rates. Centers for Disease Control and Prevention. Published August 3, 2021. Accessed April 3, 2022. https://www.cdc.gov/breastfeeding/data/nis_data/results.html 24. Bookhart LH, Joyner AB, Lee K, Worrell N, Jamieson DJ, Young MF. Moving Beyond Breastfeeding Initiation: A Qualitative Study Unpacking Factors That Influence Infant Feeding at Hospital Discharge Among Urban, Socioeconomically Disadvantaged Women. Journal of the Academy of Nutrition and Dietetics. 2021;121(9):1704-1720. doi:10.1016/j.jand.2021.02.005 25. Snyder K, Hulse E, Dingman H, Cantrell A, Hanson C, Dinkel D. Examining supports and barriers to breastfeeding through a socio-ecological lens: a qualitative study. International Breastfeeding Journal. 2021;16(1):52. doi:10.1186/s13006-021-00401-4 26. Breastfeeding Expert Work Group. Barriers to Breastfeeding: Supporting Initiation and Continuation of Breastfeeding. ACOG. Published online 2021. Accessed March 28, 2022. https://www.acog.org/en/clinical/clinical-guidance/committee-opinion/articles/2021/02/barriers-to- breastfeeding-supporting-initiation-and-continuation-of-breastfeeding 40 27. Gross RS, Mendelsohn AL, Arana MM, Messito MJ. Food Insecurity During Pregnancy and Breastfeeding by Low-Income Hispanic Mothers. Pediatrics. 2019;143(6). doi:10.1542/peds.2018- 4113 28. Dinour LM, Rivera Rodas EI, Amutah-Onukagha NN, Doamekpor LA. The role of prenatal food insecurity on breastfeeding behaviors: findings from the United States pregnancy risk assessment monitoring system. International Breastfeeding Journal. 2020;15(1):30. doi:10.1186/s13006-020- 00276-x 29. Orozco J, Echeverria SE, Armah SM, Dharod JM. Household Food Insecurity, Breastfeeding, and Related Feeding Practices in US Infants and Toddlers: Results From NHANES 2009–2014. Journal of Nutrition Education and Behavior. 2020;52(6):588-594. doi:10.1016/j.jneb.2020.02.011 30. Park K, Kersey M, Geppert J, Story M, Cutts D, Himes JH. Household food insecurity is a risk factor for iron-deficiency anaemia in a multi-ethnic, low-income sample of infants and toddlers. Public Health Nutr. 2009;12(11):2120-2128. doi:10.1017/S1368980009005540 31. Orr SK, Dachner N, Frank L, Tarasuk V. Relation between household food insecurity and breastfeeding in Canada. CMAJ. 2018;190(11):E312-E319. doi:10.1503/cmaj.170880 32. Frazier CM, Dharod J, Labban J, et al. Breastfeeding: How is it related to food insecurity and other factors among low-income mothers? Health Care Women Int. Published online July 19, 2021:1-12. doi:10.1080/07399332.2021.1929992 33. McIsaac KE, Stock DC, Lou W. Household food security and breast-feeding duration among Canadian Inuit. Public Health Nutr. 2017;20(1):64-71. doi:10.1017/S136898001600166X 34. McIsaac KE, Sellen DW, Lou W, Young K. Prevalence and Characteristics Associated with Breastfeeding Initiation Among Canadian Inuit from the 2007-2008 Nunavut Inuit Child Health Survey. MATERNAL AND CHILD HEALTH JOURNAL. 2015;19(9):2003-2011. doi:10.1007/s10995-015-1712-3 35. McIsaac KE, Lou W, Sellen D, Young TK. Exclusive Breastfeeding among Canadian Inuit: Results from the Nunavut Inuit Child Health Survey. JOURNAL OF HUMAN LACTATION. 2014;30(2):229-241. doi:10.1177/0890334413515752 36. Orr CJ, Ben-Davies M, Ravanbakht SN, et al. Parental Feeding Beliefs and Practices and Household Food Insecurity in Infancy. Acad Pediatr. 2019;19(1):80-89. doi:10.1016/j.acap.2018.09.007 37. Brown LS, Colchamiro R, Edelstein S, Metallinos-Katsaras E. Effect of Prenatal and Postpartum Food Security Status on Breastfeeding Initiation and Duration in Massachusetts WIC Participants 2001–2009. The FASEB Journal. 2013;27(S1):1054.13-1054.13. doi:10.1096/fasebj.27.1_supplement.1054.13 38. Gross RS, Mendelsohn AL, Messito MJ. Additive effects of household food insecurity during pregnancy and infancy on maternal infant feeding styles and practices. Appetite. 2018;130:20-28. doi:10.1016/j.appet.2018.07.016 39. United States Department of Agriculture. QUESTIONNAIRE - DECEMBER 2020 FOOD SECURITY SUPPLEMENT. Published online 2020. 41 40. CDC. About Adult BMI. Centers for Disease Control and Prevention. Published March 17, 2022. Accessed April 27, 2022. https://www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/index.html 41. Developing a common metric for depression across adulthood: Linking PROMIS depression with the Edinburgh Postnatal Depression Scale. - PsycNET. Accessed February 24, 2022. https://content.apa.org/record/2021-49349-001 42. Levis B, Negeri Z, Sun Y, Benedetti A, Thombs BD, DEPRESsion Screening Data (DEPRESSD) EPDS Group. Accuracy of the Edinburgh Postnatal Depression Scale (EPDS) for screening to detect major depression among pregnant and postpartum women: systematic review and meta-analysis of individual participant data. BMJ. 2020;371:m4022. doi:10.1136/bmj.m4022 43. CDC. Preterm Birth. Centers for Disease Control and Prevention. Published November 1, 2021. Accessed April 27, 2022. https://www.cdc.gov/reproductivehealth/maternalinfanthealth/pretermbirth.htm 44. Hughes MM, Black RE, Katz J. 2500-g Low Birth Weight Cutoff: History and Implications for Future Research and Policy. Matern Child Health J. 2017;21(2):283-289. doi:10.1007/s10995-016- 2131-9 45. Keddem S, Frasso R, Dichter M, Hanlon A. The Association Between Pregnancy Intention and Breastfeeding. J Hum Lact. 2018;34(1):97-105. doi:10.1177/0890334417725032 46. Collins R. Does Pregnancy Intent Impact the Decision to Breastfeed? Undergraduate Honors Theses. East Tennessee State University; 2012. 47. Napierala M, Mazela J, Merritt TA, Florek E. Tobacco smoking and breastfeeding: Effect on the lactation process, breast milk composition and infant development. A critical review. Environmental Research. 2016;151:321-338. doi:10.1016/j.envres.2016.08.002 48. Goldade K, Nichter M, Nichter M, Adrian S, Tesler L, Muramoto M. Breastfeeding and Smoking among Low-Income Women: Results of a Longitudinal Qualitative Study. Birth. 2008;35(3):230- 240. doi:10.1111/j.1523-536X.2008.00244.x 49. Edelblute HB, Altman CE. The Interaction and Impact of Social Support and Father Absence on Breastfeeding. Breastfeed Med. 2021;16(8):629-634. doi:10.1089/bfm.2020.0202 50. Corsack C, Wallenborn JT, Harley KG, Eskenazi B. Parental Cohabitation and Breastfeeding Outcomes Among United States Adolescent Mothers. Breastfeeding Medicine. 2022;17(1):72-78. doi:10.1089/bfm.2021.0090 51. Heck KE, Braveman P, Cubbin C, Chávez GF, Kiely JL. Socioeconomic Status and Breastfeeding Initiation Among California Mothers. Public Health Rep. 2006;121(1):51-59. 52. Gaffney KF, Brito AV, Kitsantas P, Kermer DA, Pereddo G, Ramos KM. Institute of Medicine Early Infant Feeding Recommendations for Childhood Obesity Prevention: Implementation by Immigrant Mothers From Central America. J Pediatr Nurs. 2018;40:27-33. doi:10.1016/j.pedn.2018.02.017 53. Alaimo K, Olson CM, Frongillo EA. Food insufficiency and American school-aged children’s cognitive, academic, and psychosocial development. Pediatrics. 2001;108(1):44-53. 42 54. Centers for Disease Control and Prevention. Rates of Any and Exclusive Breastfeeding by Sociodemographics among Children Born in 201.; 2020. Accessed August 11, 2021. https://www.cdc.gov/breastfeeding/data/nis_data/rates-any-exclusive-bf-socio-dem-2012.htm 55. Zhang Q, Lamichhane R, Wright M, McLaughlin PW, Stacy B. Trends in Breastfeeding Disparities in US Infants by WIC Eligibility and Participation. J Nutr Educ Behav. 2019;51(2):182-189. doi:10.1016/j.jneb.2018.10.005 56. Amir LH, Donath S. A systematic review of maternal obesity and breastfeeding intention, initiation and duration. BMC Pregnancy Childbirth. 2007;7(1):9. doi:10.1186/1471-2393-7-9 57. Hashemi-Nazari SS, Hasani J, Izadi N, et al. The effect of pre-pregnancy body mass index on breastfeeding initiation, intention and duration: A systematic review and dose-response meta- analysis. Heliyon. 2020;6(12):e05622. doi:10.1016/j.heliyon.2020.e05622 58. Perez MR, Castro LS de, Chang YS, et al. Breastfeeding Practices and Problems Among Obese Women Compared with Nonobese Women in a Brazilian Hospital. Women’s Health Reports. 2021;2(1):219-226. doi:10.1089/whr.2021.0021 59. Lyons S, Currie S, Smith DM. Learning from Women with a Body Mass Index (Bmi) ≥ 30 kg/m2 who have Breastfed and/or are Breastfeeding: a Qualitative Interview Study. Matern Child Health J. 2019;23(5):648-656. doi:10.1007/s10995-018-2679-7 60. Mercier RJ, Burcher TA, Horowitz R, Wolf A. Differences in Breastfeeding Among Medicaid and Commercially Insured Patients: A Retrospective Cohort Study. Breastfeeding Medicine. 2018;13(4):286-291. doi:10.1089/bfm.2017.0228 61. Lazar M, Davenport L. Barriers to Health Care Access for Low Income Families: A Review of Literature. Journal of Community Health Nursing. 2018;35(1):28-37. doi:10.1080/07370016.2018.1404832 62. Gattu RK, Paik G, Wang Y, Ray P, Lichenstein R, Black MM. The Hunger Vital Sign Identifies Household Food Insecurity among Children in Emergency Departments and Primary Care. Children (Basel). 2019;6(10):E107. doi:10.3390/children6100107 63. Garg A, Toy S, Tripodis Y, Silverstein M, Freeman E. Addressing Social Determinants of Health at Well Child Care Visits: A Cluster RCT. Pediatrics. 2015;135(2):e296-e304. doi:10.1542/peds.2014-2888 64. Makelarski JA, Abramsohn E, Benjamin JH, Du S, Lindau ST. Diagnostic Accuracy of Two Food Insecurity Screeners Recommended for Use in Health Care Settings. Am J Public Health. 2017;107(11):1812-1817. doi:10.2105/AJPH.2017.304033 65. Rose-Jacobs R, Trevino-Talbot M, Lloyd-Travaglini C, et al. Could prenatal food insecurity influence neonatal abstinence syndrome severity? Addiction. 2019;114(2):337-343. doi:10.1111/add.14458 66. Metallinos-Katsaras E, Gorman KS, Wilde P, Kallio J. A longitudinal study of WIC participation on household food insecurity. Matern Child Health J. 2011;15(5):627-633. doi:10.1007/s10995-010- 0616-5 43