Objective. To determine the prevalence and correlates of the early discontinuation of breastfeeding by mothers eligible for the Women, Infants, and Children Program (WIC).
Methodology. A longitudinal observational study in which we enrolled English-speaking mothers who initiated breastfeeding after delivering healthy-term infants at Yale-New Haven Hospital and planned to bring their infants to the hospital's primary care center. Data on mother's baseline knowledge, attitudes, beliefs, and problems regarding breastfeeding were collected by semistructured interviews within 48 hours after delivery, at 1 and 2 weeks' postpartum, and by chart reviews at 2 and 4 months. A nonparticipating control group was used to test the Hawthorne effect.
Results. Of the 64 participating mothers, the majority were minority (56% black, 34% of Puerto Rican origin), single (75%), and already enrolled in WIC (91%). The rates of discontinuation of breastfeeding were 27%, 37%, 70%, and 89% by 1 week, 2 weeks, 2 months, and 4 months, respectively. The mother's knowledge and problems of lactation were not associated with the early discontinuation of breastfeeding. After using logistic regression to control for potential confounders, mothers who lacked confidence at baseline that they would still be breastfeeding at 2 months (risk ratio: 2.38, 95% confidence interval: 1.82–6.18), and those who believed that the baby prefers formula (risk ratio: 1.68, 95% confidence interval: 1.04–2.71) were more likely to stop breastfeeding within the first 2 weeks postpartum.
Conclusions. The results of this study demonstrate that interventions aimed at prolonging the duration of breastfeeding in this population will need to shift focus from increasing knowledge and managing problems of lactation to enhancing the mother's confidence regarding breastfeeding, while also addressing beliefs regarding an infant's preferences.
Breastfeeding has been key for the survival of the human species1 and has been a normative behavior for all populations until the advent of formula. The widespread behavior of nursing has preceded the knowledge of the benefits of breastfeeding and advantages over formula. Lack of knowledge regarding breastfeeding and problems of lactation may be universally encountered. These factors, however, play a minor role in the discontinuation of breastfeeding in populations around the world that practice prolonged breastfeeding.2–5 Knowledge of when and why women discontinue breastfeeding is crucial to guide interventions aimed at increasing both the initiation and duration of breastfeeding. The longer women continue nursing, the more this method of infant feeding will become “seen” and the more likely it is to be viewed as normative behavior, rather than just a “try.”
The goals of the Department of Health and Human Services Healthy People 2000 objective regarding breastfeeding were 75% for the initiation of breastfeeding, and 50% for the continuation up to 6 months postpartum.6 Although in the United States from 1989 to 1995, the initiation of breastfeeding increased from 52.2% to 59.7% and the rate of breastfeeding at 6 months increased from 18.1% to 21.6%, these rates are lower for more indigent women.7Many interventions have intended to increase the rates of both the initiation and continuation of breastfeeding for women enrolled in the Women, Infants, and Children Program (WIC) with limited success.8–14 Most interventions have focused on increasing knowledge regarding breastfeeding or the management of problems of lactation. Where breastfeeding is not the norm, however, it is not known whether the lack of knowledge and problems encountered during breastfeeding are the reasons for the early discontinuation of breastfeeding.
In the United States, for low-income mothers, regional differences seem to play a role in the rates of the initiation and duration of breastfeeding. Although there are some data about the correlates of breastfeeding for rural populations11–15 and for women living in Southern,16–20 Western,21,22 and Midwestern23 United States, specifically for mothers living in urban environments and enrolled in WIC, recent data about the correlates of breastfeeding duration in the Northeastern region are few.10 Furthermore, although data on the predictors of the early discontinuation of breastfeeding by higher income women are present,24 the precise timing and predictors for the early discontinuation of breastfeeding are not available for low-income populations.
Therefore, the purposes of this study were to identify the rate, timing, and predictors of the early discontinuation of breastfeeding for women enrolled in WIC and living in an urban Northeastern environment. Based on the theory by Fishbein and Ajzen25,26 that behavioral intentions are the most important predictors of behavior and on previous reports that have shown that intentions related to breastfeeding predict breastfeeding behavior,27–29 we hypothesized that for those women living in impoverished urban environments 1) the early discontinuation of breastfeeding would not be dependent on the lack of knowledge on breastfeeding or on problems of lactation; but that 2) the early discontinuation of breastfeeding could be predicted during the early postpartum period by the mother's beliefs and attitudes, specifically, her report of a lack of confidence in her intention to continue breastfeeding.
During a 6-month period, we reviewed the daily postpartum logs at Yale-New Haven Hospital (a university-based hospital) serving an inner-city population and enrolled mothers who met the following criteria: 1) spoke English and were eligible for WIC; 2) delivered a healthy-term, singleton infant; 3) planned to bring their infant to the hospital's primary care center for well child care; and 4) initiated breastfeeding during the first 48 hours postpartum.
Mothers were told that we were interested in learning about breastfeeding in general, and oral consent to participate in the study was obtained. The study was approved by the Institutional Review Board of Yale University School of Medicine. Baseline data were collected within the first 48 hours after delivery through a semistructured interview conducted on the postpartum floor. This interview was developed specifically for the study and included both closed-ended and precoded, open-ended questions containing information on factors related to the discontinuation of breastfeeding. This information included the mother's: 1) sociodemographic characteristics; 2) knowledge regarding breastfeeding; and 3) attitudes and beliefs related to breastfeeding; in particular, her confidence in her intention to continue breastfeeding. A mother was considered not confident if she stated at the baseline interview that her probability of still breastfeeding at 2 months postpartum was low.
Problems related to breastfeeding and outcome data on the method of feeding were collected by a telephone interview conducted 1-week postpartum, an interview in person conducted at the 2-week well-child visit, and a review of medical records at 2 and 4 months of the child's life. If a mother stated that she was no longer breastfeeding at all, this was considered discontinuation; any other amount of breastfeeding was considered continuation. The discontinuation of all breastfeeding episodes by 2 weeks or 2 months was referred to as the “very early” or “early” discontinuation of breastfeeding, respectively. Exclusive breastfeeding was defined as breastfeeding without the introduction of formula or solids. Water supplementation was allowed in the definition of exclusive breastfeeding.
Because data collection involved intensive interviewing and patient contact, we also examined the possibility of the Hawthorne effect. This refers to the effect of study participation on the outcome.30 A nonparticipating control group (NPCG) was used to test the Hawthorne effect. This group consisted of women who fulfilled the same eligibility criteria but who were purposefully not interviewed. The researchers conducted interviews with study group mothers on alternating days. Women for the NPCG were chosen to be those mothers who delivered on days that the researchers did not conduct interviews. Baseline sociodemographic data and similar outcome data regarding feeding practices were collected for this group by medical chart review.
The rates of discontinuing breastfeeding were calculated by dividing the number of women not breastfeeding at all by the number of women present in the study at each outcome point. Relative risks and confidence intervals (CIs) were computed to examine the relationships between the baseline variables, interim problems related to breastfeeding, and the outcome of breastfeeding at 2 weeks and 2 months separately. Those variables significantly related to the outcome were entered into a step-wise logistic regression model to identify variables that were independently related to the discontinuation of breastfeeding at 2 weeks. Because of the low rate of continuation at 2 months, a logistic regression analysis was not performed on this outcome. Because of the very low rate of breastfeeding at 4 months, these data were used only to provide rates of discontinuation at the 2- to 4-month interval. For all analyses, we used SAS-PC (SAS Institute Inc, Cary, NC).31
During the time of data collection, information on 457 mothers was reviewed prospectively. Of these, 125 (27.4%) initiated breastfeeding and 332 (72.6%) bottle-fed their infants during the first 48 hours postpartum. Of the 125 breastfeeding mothers, 64 (51.2%) were enrolled in the study group, and 61 (48.8%) were enrolled in the NPCG. Of the 61 mothers enrolled in the NPCG, 21 (34.4%) stopped breastfeeding by 2 weeks and 72.7% and 83.6% had stopped breastfeeding by 2 and 4 months, respectively. There were no differences between the study group and the NPCG in any of the baseline sociodemographic characteristics that were noted in the medical records or in the outcome of breastfeeding. These results showed that participation in this study did not change the rates of breastfeeding. The remainder of the results focuses on the study group.
Sociodemographic Characteristics of Sample
As seen in Table 1, most mothers were of black (56.3%) or Puerto Rican (34.4%) origin, and most were single (75.0%); their median age was 22.0, and median years of education was 12.1. The majority of mothers (90.6%) had already been enrolled in WIC before delivery. Most mothers (71.9%) stated that they would not be going back to work or school within the next 2 months.
Knowledge Regarding Breastfeeding
The primary source of information on breastfeeding was identified as written material or videotapes by 49.2% of mothers, medical professionals (including obstetrician, midwife, pediatrician, or nurse) by 22.1%, and family or friends by 16.9%; 11.9% could not name a source, and only 1 mother identified a WIC consultant as a source of information. The package of benefits provided by WIC to breastfeeding mothers was correctly known by only 3 mothers. Almost all mothers (92.3%) noted that the infant's pediatrician was the primary source from whom they would desire to receive information regarding breastfeeding, and the pediatrician was also identified by 79.7% as the primary person to turn to, if problems related to breastfeeding should occur. Most mothers were knowledgeable about the benefits of breastfeeding for the infant (90.2%), naming all 3 types of benefits—nutritional, immunologic, and psychological. Knowledge related to the practice of breastfeeding, such as how long exclusive breastfeeding is sufficient for the baby, was correctly known by 30% of the women only, and only 3 mothers identified frequent nursing as a means of increasing milk production.
Attitudes and Beliefs Related to Breastfeeding
Primiparous mothers comprised 51.6% of the study group, and of the 31 multiparous mothers, 14 (42.4%) had not breastfed a former child. Thirty-one percent of mothers had not seen anyone nursing, and 51.6% had not been breastfed themselves. Benefits for the infant, such as protection against illness, best food, best nutrition, and no artificial products, were stated most frequently as the primary reason to breastfeed (92.2%). Enhancement of mother-infant interactions (3.1%) or benefits for the mother (4.7%) were stated less often. The decision to breastfeed was made before pregnancy by 32.8% of the mothers and during the first, second, and third trimesters by 46.9%, 12.5%, and 7.5% of the mothers, respectively. When asked how mothers thought people close to them felt about their decision to breastfeed, 70.3% stated that such individuals wanted them to breastfeed. In 62.5% of the cases, the father of the baby was perceived as supportive of breastfeeding.
Most mothers did not believe that breastfeeding was inconvenient for the mother (81.3%); however, 57.8% believed it was easier than formula feeding, 28.1% felt comfortable nursing in public, and 32.8% thought it was in fashion. Although most mothers stated that babies enjoy breastfeeding more than formula feeding, 39.1% stated that babies enjoy formula more. Most mothers (95.3%) stated that they wanted to breastfeed up to 6 months, and 60.9% stated that they wanted to breastfeed exclusively. Their intended reasons for stopping were related to a natural weaning time indicated by comments like “when baby gets teeth,” or “when he is old enough.” When asked about their confidence that they would continue breastfeeding until the infant was 2 months of age, 45.3% stated that the chances that they would still be breastfeeding were low.
Problems Experienced During the First 2 Weeks of Breastfeeding
At the 2-week interview, most mothers reported having experienced problems related to breastfeeding. These were classified into 5 groups: 1) problems of nursing, such as cracked and painful nipples, soreness, and leaking, were reported by 26.6% of the mothers; 2) problems related to the baby, such as spitting up, crankiness, and fussing, were reported by 35.9%; 3) a perception of insufficient milk supply was reported by 28.1%; 4) inconvenience for the mother, such as being too tiring or preventing the mother from going out, were reported by 28%; and 5) problems that required medical management, such as jaundice or maternal illness, were reported by 10.9%.
Outcome of Breastfeeding
All 64 mothers participated in the 1-week and 2-week interviews. Feeding methods could be documented for 53 (82.8%) and 51 (79.7%) infants at 2 and 4 months, respectively. There were no differences between the 64 mothers or infants enrolled in the study and those 13 lost to follow-up by 4 months, with regard to baseline variables or rates of discontinuation of breastfeeding at 2 weeks. Attrition was attributable to missing charts or missing data on feeding in the charts. As seen in Table 2, over a 4-month period, there were 2 time frames during which mothers most frequently discontinued breastfeeding. The first peak of discontinuation occurred during the first week, when 26.6% mothers discontinued breastfeeding; the second peak was between 2 weeks and 2 months, when another 32.1% stopped. By 2 months and 4 months, 77.4% and 88.2% of mothers had stopped breastfeeding, respectively. In addition to these low rates of continuation of any amount of breastfeeding, as seen in Table 2, exclusive breastfeeding rates were similarly low and decreased in time. At 1 week, 2 weeks, 2 months, and 4 months: 37.5%, 26.6%, 11.3%, and 5.9% of the total number of infants remaining in the study were exclusively breastfed, respectively.
Factors Predicting the Early Discontinuation of Breastfeeding
Table 3 shows the relative risks for those baseline factors that were significantly related to the early discontinuation of breastfeeding at 2 weeks or 2 months. Although several factors were associated with early discontinuation, only 2 variables—maternal age <20 and lack of confidence about continuing to breastfeed until the infant was 2 months of age—were significantly associated with discontinuing breastfeeding both at 2 weeks and at 2 months.
Sources of information and knowledge on breastfeeding or the occurrence of breastfeeding problems during the first 2 weeks were not related to the outcome of discontinuation. Mothers who had not breastfed before were more likely to stop breastfeeding very early, but this did not reach a level of significance (risk ratio [RR]: 2.53; 95% confidence interval [CI]: 0.86–7.42). Attitudes and beliefs that were not related to the discontinuation of breastfeeding included perceived support regarding breastfeeding, intended duration of breastfeeding, intended exclusion of formula, and beliefs of the convenience or hardship of breastfeeding.
Independent Variables Predicting the Early Discontinuation of Breastfeeding
To determine the independent predictors of the very early discontinuation of breastfeeding, a step-wise logistic regression analysis model was used. The 7 baseline variables shown in Table 3 were entered into the model as independent variables, with the termination of breastfeeding at 2 weeks as the dependent variable. Two independent predictors of breastfeeding at 2 weeks emerged from the logistic regression model: the mother's lack of confidence about continuing to breastfeed up to 2 months (Mantel-Haenszel adjusted RR = 2.38, 95% CI: 1.82–6.18), and the belief that the baby enjoys bottle feeding (Mantel-Haenszel adjusted RR = 1.68, 95% CI: 1.04–2.71). As seen previously in Table 3, of these 2 factors, only confidence in the continuation of breastfeeding was significantly related to that outcome at 2 months. Those mothers who were not confident that they would continue nursing up to 2 months were almost 12 times more likely to stop breastfeeding before 2 months than those who were confident.
This study provides a longitudinal description of the rate, timing, and factors affecting the very early termination of breastfeeding by mothers eligible for WIC and living in an urban area. In our study, 27% of English speaking mothers who were eligible for WIC and who delivered healthy-term, singleton infants initiated breastfeeding. This rate is consistent with other reports from the Northeast region10,32 and also rates from urban populations from other regions.15,16 Although rates of the early discontinuation of breastfeeding have been investigated less intensively, our rate of discontinuation at 2 weeks (37.4%) is similar to that reported by Kistin32 (38.5%), and our rate of discontinuation at 2 months (77.4%) is similar to that reported by Serwint (72.7%).33 These similar rates of discontinuation justify the generalizability of our results and the need to understand better when and why women in low-income populations discontinue breastfeeding early.
The exact timing of the termination of breastfeeding has not been previously reported. We found that during the first 4 months postpartum, there are 2 peaks for the termination of breastfeeding. The first peak is during the first week postpartum when approximately 25% of mothers stop breastfeeding, and the second peak is between 2 weeks and 2 months postpartum when another 40% discontinue. Fewer mothers (approximately 10%) stop breastfeeding between 1 and 2 weeks or between 2 months and 4 months. This timing suggests that a 1-week postpartum visit for well-child care is too late to intervene for many breastfeeding mothers.
Our first hypothesis that the early discontinuation of breastfeeding would not be dependent on problems of lactation or on a lack of knowledge about breastfeeding was confirmed. Although the majority of mothers in our sample reported experiencing problems, there was no significant relationship between the very early discontinuation of breastfeeding and having problems related to breastfeeding. This result may help explain why interventions aimed at solving problems of nursing have reported limited success.10,34
The perception of insufficient milk supply was investigated separately, because parallel to other investigators35,36 we believed that this would be an important reason why mothers discontinue breastfeeding. Contrary to our expectations, however, the perception of insufficient milk supply was not related to the early termination of breastfeeding.
Most mothers were knowledgeable about the benefits of breastfeeding. They extensively described the benefits for the baby, and these were the primary reasons given by mothers for the initiation of breastfeeding. Information related to the actual practice of breastfeeding was less often known by the mothers. Although the majority of mothers had decided to breastfeed before or during the first trimester of pregnancy, and although the majority had been enrolled in WIC during the first 2 trimesters of pregnancy (79.3%), the benefits provided by the WIC program to breastfeeding mothers were primarily unknown. Overall, there was little variation in information that was known, and there was no significant relationship between knowledge regarding breastfeeding and the early termination of nursing. This finding has implications for interventions that aim to increase the duration of breastfeeding. Although information that is already known by most mothers—namely, the benefits of breastfeeding—may play a very important role in the initiation of breastfeeding, it does not affect the duration of breastfeeding.
As indicated by a comprehensive review of the literature, sociodemographic variables have been consistently related to the initiation and discontinuation of breastfeeding.37 We found younger age to be significantly related to the discontinuation of breastfeeding by 2 weeks and 2 months. Mothers who had less than a high school education and those of Puerto Rican origin had higher rates of discontinuation of breastfeeding during the first 2 weeks postpartum. These findings regarding mothers of Puerto Rican ethnicity are different from those reported for other Hispanic mothers.38 Often in the medical literature, populations of Mexican and Puerto Rican origin are both categorized as Hispanic. Our findings support the need to study additionally the breastfeeding patterns of mothers of Puerto Rican origin living in the United States, to define ethnicity accurately, and also to guide interventions specifically to this ethnic group, particularly during the earlier phases of breastfeeding.
Perceived support from significant others, a factor often found to be related to the initiation of breastfeeding,39,40 was not confirmed in our study. We may not have been able to determine actual support from our interview. Alternatively, our finding, when coupled with results from other studies,41,42 suggests that support from the father of the baby or other persons close to the mother may not be enough or crucial for the continuation of breastfeeding.
Our results supported our second hypothesis that confidence in the intention to breastfeed would be the strongest predictor of breastfeeding outcome. The discontinuation of breastfeeding could be predicted by the mother's own prediction shortly after birth that she would not likely breastfeed for long. This information, which was obtained before the mother had actually experienced breastfeeding for long or before she had encountered problems, suggests strongly that the discontinuation of breastfeeding is determined very early. Our findings are similar to those of Loughlin et al in a more affluent population of mostly white mothers followed in a private practice.24These mothers also were asked how confident they were in breastfeeding, and this variable emerged as a significant predictor of the early termination of breastfeeding. Interestingly, in our study, when asked about how long they wished to continue breastfeeding, the vast majority stated that they would continue until the baby was old enough to wean. It was only when mothers were asked about their confidence or certainty that they would continue that mothers were able to express their doubts and lack of confidence.
Based on our results, we believe that the concept of confidence in breastfeeding should be addressed in any program aimed at promoting breastfeeding, and the mother should be allowed open discussion of this doubt or confidence. This method of intervention may help a woman to find the reasons behind the discrepancies between her idealized desire to continue breastfeeding and her lack of confidence that she actually will continue. Understanding these discrepancies, in turn, may help to increase her confidence in her intended behavior. We believe that for populations where breastfeeding is not a normative behavior, questions related to why it is not and why people have doubt should be posed early in life, such as in school programs or through mass media, or at least very early in pregnancy. As implied earlier in the literature,43 another approach that might strengthen confidence may be the provision of financial incentives such as additional vouchers for mothers who choose to breastfeed coupled with support to increase confidence.
A mother's belief that her baby enjoys formula feeding more than breastfeeding emerged as a second independent predictor of the very early discontinuation of breastfeeding. Rephrased, this new finding implies that the mother's confidence in the infant's confidence that he/she wanted breastfeeding seemed as important as the mother's confidence in herself that she would continue to breastfeed. Although these 2 concepts seem to be related and overlapping, in our study, they were only moderately correlated (Pearson's R = 0.30) and emerged as independent predictors of breastfeeding at 2 weeks. In the study by Loughlin,24 the postpartum nurse's ratings of the newborn's feeding behaviors also were predictive of the breastfeeding duration. It may be that a newborn's early behavior at the breast may shape a mother's perception of whether her baby enjoys breastfeeding. It also may be, as implied by Loughlin, that the mother-infant interaction, which may be guided by the mother's confidence, is shaping the nurse's ratings. Our findings suggest that the mother's perception of the infant's response to breastfeeding is separate from the mother's confidence in breastfeeding. These preliminary findings highlight that the infant is an active participant in the outcome of breastfeeding. The infant's cues and what the infant wants need to be discussed to dispel the myth that infants like formula better.
Mothers identified their pediatrician as the desired source of information regarding breastfeeding and problems with breastfeeding. These findings place pediatricians serving low-income, urban populations in a key position to promote breastfeeding. Recent information regarding the importance of a physician's confidence about breastfeeding counseling44 and the educational needs of pediatricians about breastfeeding45 suggests ways to influence how physicians promote breastfeeding.
Our results reflect mothers enrolled in WIC at 1 center, therefore, the generalizability is limited. Furthermore, because the rate of continuation is low, a greater number of participants are needed to examine the predictors of continuation of breastfeeding beyond 2 months. Interviews in person would have been desirable to obtain information about breastfeeding at 2 and 4 months. Because of limitations of resources for such a follow-up, we relied on chart reviews. Despite these limitations, we believe that we have been able to determine attitudinal factors associated with the duration of breastfeeding, and that our results add to the literature concerning the sociodemographic characteristic. Attitudinal factors and beliefs involve intensive interviewing; we believe that the inclusion of a NPCG to test the Hawthorne effect has strengthened the validity of our results.
We have demonstrated that breastfeeding women in an underserved population make up their minds to breastfeed before conventional prenatal breastfeeding promotion programs can reach them, and that usually such women have made up their minds previously to stop breastfeeding sooner than conventional well-child care programs can reach them. We conclude that to make breastfeeding the norm for a population where this is an infrequent event, interventions should focus on increasing a mother's confidence in breastfeeding; in addition, the role of the infant on the duration of breastfeeding should be taken into account.
Dr. Leventhal was supported in part by a Training Grant in Behavioral Pediatrics from the Bureau of Maternal and Child Health.
- Received December 2, 1999.
- Accepted December 2, 1999.
Reprint requests to (I.O.E.) Assistant Professor of Pediatrics, Department of Pediatrics, Ankara University School of Medicine, Cebeci, 06100, Ankara, Turkey. E-mail:
- WIC =
- Women, Infants, and Children Program •
- NPCG =
- nonparticipating control group •
- CI =
- confidence interval •
- RR =
- risk ratio
- Jakobsen MS,
- Sodemann M,
- Molbak K,
- Aaby P
- US Department of Health and Human Services. Healthy People 2000: National Health Promotion and Disease Prevention Objectives-Full Report, with Commentary. Washington DC: US Government Printing Office. 2000;1990:379
- Ryan AS. The resurgence of breast feeding in the United States.Pediatrics. 1997;99(4). URL:www.pediatrics.org/cgi/content/full/99/4/e12
- Caulfield LE,
- Gross SM,
- Bentley ME,
- et al.
- Reifsner E,
- Eckhart D
- Long DG,
- Funk Archuleta MA,
- Geiger CJ,
- Mozar AJ,
- Heins JN
- Gielen AC,
- Faden RR,
- O'Campo P,
- Paige DM
- MacGowan RJ,
- MacGowan CA,
- Serdula MK,
- Lane JM,
- Joesoef RM,
- Cook FH
- Bagwell JE,
- Kendrick OW,
- Stitt KR,
- Leeper JD,
- Espy ML,
- Gedel ML
- McClurg-Hitt D,
- Olsen J
- Loughlin HH,
- Clapp-Channing NE,
- Gehlbach SH,
- Pollard JC,
- McCutchen TM
- Homans G. Group Factors in Worker Productivity. In: H Proshansky, B Seidenberg, eds. Basic Studies in Social Psychology. New York, NY: Holt, Rinehart, and Winston; 1965:592–604
- SAS System for Elementary Statistical Analysis. 2nd ed. Cary, NC: SAS Institute Inc; 1997
- Kistin N,
- Benton D,
- Rao S,
- Sullivan M
- Serwint JR,
- Wilson ME,
- Vogelhut JW,
- Repke JT,
- Seidel HM
- Losch M,
- Dungy Claibourne ID,
- Russell D,
- Dusdieker LB
- Ryan AS,
- Rush D,
- Krieger FW,
- et al.
- Raj VK
- Kramer MS
- Schanler RJ, O'Connor KG, Lawrence RA. Pediatricians practices and attitudes regarding breast feeding promotion. Pediatrics. 1999; 103(3). URL:www.pediatrics.org/cgi/content/full/103/3/e35
- Copyright © 2001 American Academy of Pediatrics