Martin C. Mahoney, MD, PhD

Healthy People 2000, a national agenda for attention to health promotion and disease-prevention activities, has established a target that 75% of mothers will breastfeed at the time of their postpartum hospital discharge and 50% will continue to breastfeed through 6 months.¹ Breastfeeding offers several advantages, including the provision of a convenient balanced nutritional source, promotion of bonding between mother and infant, and conferral of an immunologic advantage because breastfed infants are less prone to respiratory and enteric infections.^2,3 Breastfeeding has also been reported to contribute to infant cognitive development and to decrease the risk of chronic diseases during childhood.⁴ Despite these documented benefits, the proportion of women who currently breastfeed their newborns remains suboptimal at 50%.¹

Breastfeeding practices vary by ethnicity, education, socioeconomic status, and maternal age.^5,6 The lowest rates are observed among young and undereducated mothers, as well as among African American mothers.^6,7 Among low-income groups, breastfeeding is initiated by approximately one third of mothers and is continued for 6 months by only 9%.⁷ Since low-income, minority populations commonly experience a higher prevalence of various infant health problems, breastfeeding seems especially important in benefiting infants from these communities.

Factors that influence breastfeeding have not been sufficiently studied among urban, low-income populations seen in family medicine centers. In an attempt to expand our knowledge of infant feeding practices among indigent populations, this research explored correlates of planned breastfeeding among postpartum women in a low-income, urban family practice setting.

Methods

This research project used a structured survey instrument to assess anticipated infant feeding practices among newly postpartum women from an urban family medicine site. This cross-sectional study was approved by the hospital Institutional Review Board. Informed consent was obtained from all participants.

Study Population

This study was completed among patients from an academic family medicine training site in Upstate New York. This facility-the Family Medicine Center (FMC), a community-based residency practice located in an urbanized area-offers a variety of medical services, including provision of comprehensive and longitudinal ambulatory family medical care. The FMC is situated in a health professional shortage area; a significant proportion of patients seen at the facility can be represented as economically disadvantaged and minority (eg, African American).

Study participants included postpartum women from the FMC who gave birth between January 1996 and June 1997. There were no restrictions based on race or age. Nearly 84% (67/80) of all eligible postpartum patients were invited to participate. No systematic patterns were identified among the subset of postpartum women (n=13) who did not receive survey forms, and no systematic bias was identified among the postpartum women not interviewed. Surveys were completed for 98.5% (66/67) of the patients asked to participate. Interviews were conducted while subjects were postpartum in-patients at an affiliated tertiary care hospital where patients from the FMC are routinely admitted for deliveries.

Survey Instrument

Using a structured survey instrument, we obtained data on demographics, relevant clinical history, and past infant feeding practices, as well as several items relating to current infant feeding plans and attitudes. Survey responses were based on both a medical chart review and an interview administered to subjects by trained individuals.

Data Analyses

Completed surveys were entered into a computerized database using SPSS/PC (SPSS/PC+, SPSS Inc., Chicago, Ill, 1990) and Epi Info (Epi Info version 5.01, Centers for Disease Control, Epidemiology Program Office, Atlanta, Ga, 1990). We used cross-tabulations and contingency table analyses to examine potential predictors of, as well as barriers to, breastfeeding. Independent variables included subject demographics, reproductive history, previous experiences with breastfeeding, and responses to attitudinal items on breastfeeding. We used odds ratios as one measure of association to assess the relationship between selected independent variables and anticipated infant breastfeeding practices. Ninety-five percent confidence intervals for the odds ratio provided an indication of both the precision and significance of these point estimates of risk. We used multiple logistic regression to calculate adjusted odds ratios and confidence intervals. Selected demographic, historical, and attitudinal variables found to be associated with breastfeeding through univariate analyses (P <.10) were entered into the logistic model.

Respondents were stratified into 2 groups on the basis of their responses to survey items on anticipated infant feeding practices. Subjects who reported plans to either exclusively breastfeed or plans to use any combination of breastfeeding and bottle-feeding were classified as the “breastfeed” group, while respondents planning to only bottle-feed their infants were classified as the “bottle-feed” group.

Results

Our results are based on interviews completed with 66 postpartum patients. The majority of participants were African American (95%) and single (95%), with a median age of 24 years. They completed a median of 12 years of schooling, had a median of 3 pregnancies, and a median 2 previous births. Selected demographic and reproductive characteristics are presented in Table 1.

Only 3 participants (4.5%) indicated that they planned to breastfeed exclusively, while an additional 11 (16.7%) noted plans to use a combination of bottle-feeding and breastfeeding. Breastfeeding plans were not found to be associated with maternal demographics features or with reproductive characteristics. A history of breastfeeding was reported by 28% of those planning to breastfeed their new infant, compared with 8% of those planning to bottle-feed; however, this difference failed to achieve significance (odds ratio [OR]=4.80; 95% confidence interval [CI], 0.74-29.7).

Analyses of attitudes about infant feeding are summarized in Table 2. Women planning to breastfeed their infants were more likely to agree that breastfeeding helped reduce their own weight after pregnancy and that they were encouraged to breastfeed by the baby’s father and by their mothers. Respondents planning to bottle-feed reported that breastfeeding was too complicated.

Multiple logistic regression yielded a model that accounted for 50% of the variance in anticipated breastfeeding practice. The adjusted model identified significant odds ratios for the association between low levels of maternal education and anticipated breastfeeding (OR=0.13; 95% CI, 0.05-0.35) and for encouragement from the baby’s father or the participant’s mother to breastfeed and anticipated breastfeeding (OR=12.4; 95% CI, 4.92-31.4). This logistic model also correctly classified 86.5% of the respondents, including 46% of the breastfeed group and 100% of the bottle-feed group. (Univariate odds ratios for these independent variables were 0.18 (0.02-0.95) for limited education and 8.60 (1.71-43.7) for encouragement.)

Discussion

The benefits of infant breastfeeding-including improved maternal-infant bonding, nutritional completeness and enhanced immunologic effects-are well documented and result in decreased infant morbidity during the first year of life.^2,3 This study contributes valuable data on anticipated infant feeding practices among predominantly African American patients from a low-income, urban community. While the goal of Healthy People 2000 is to attain a level of 75% of mothers who breastfeed in the immediate postpartum period, only 21% of study subjects indicated plans to breastfeed their infants. It should be noted that this figure likely represents an overestimate of actual infant feeding practices, since only 3 subjects (4.5%) reported plans to exclusively breastfeed their infants. Moreover, respondents planning to breastfeed noted intentions to nurse their infants for periods ranging from 4 to 52 weeks (median = 8 weeks) and for 30% to 100% of all feedings (median = 75% of feedings). The plans described by these mothers suggest a weak commitment to breastfeeding. For breastfeeding to be successful in the immediate postpartum period it is important that breast milk be the exclusive nutritional source. In addition, the plans for the limited duration of breastfeeding lag behind the public health goal to maintain a level of 50% breastfeeding at 6 months of age.

This study examined anticipated feeding, rather than actual feeding practices, to allow for data collection without the need for further follow-up after delivery. Previous research suggests that intention to breastfeed is related to both initiation^8,9 and duration¹⁰ of breastfeeding. Because, in part, of the limited numbers of study subjects planning to breastfeed their infants (n=14), many of our comparisons failed to achieve statistical significance. Using the chi-square test of independence, the observed sample sizes, and a 2-sided test with a level of significance of 5%, there was approximately 80% power to detect absolute differences of 30% to 35% in the self-reported infant feeding plans by demographic group or attitudinal response. This study design might be replicated among a larger number of postpartum patients at other family medicine centers.

Several significant associations were identified, including limited maternal educational attainment and decreased likelihood of infant breastfeeding (OR=0.18). This association between breastfeeding and education has been noted previously.^5,6 In addition, encouragement from the subject’s mother was noted to increase the likelihood of breastfeeding 12-fold. The broad application of breastfeeding promotion programs to include members of a woman’s social support system has been suggested as a means of increasing breastfeeding rates.^9,11,12 Similarly, the lack of social support for breastfeeding has been associated with increased bottle-feeding.¹³

A recent article reported limited knowledge of breastfeeding issues and techniques among family medicine attending physicians and residents.¹⁴ To address these knowledge gaps findings from this study should be used to develop educational interventions that actively promote breastfeeding among members of this community and in similar patient populations. Family medicine centers, for example, could develop a family-centered intervention whereby mothers and grandmothers of pregnant patients might use a structured forum to encourage breastfeeding by reviewing known benefits and sharing stories of successful breastfeeding experiences. The importance of soliciting and maintaining the involvement of the baby’s father in the promotion of breastfeeding is supported by our study, as well as by previous reports.^8,15-17 In this educational context, fathers might be persuaded to attend prenatal visits where the topic of infant feeding is addressed. Various financial incentives could also be used to stimulate the participation of urban, minority pregnant women and their families and partners in educational workshops.¹⁸ These forums should provide opportunities to address the prevalent misperceptions regarding the effects and benefits of breastfeeding.

Conclusions

This study presents unique data regarding anticipated infant feeding practices among predominantly African American patients from a low-income, urban community. The results reveal decreased anticipated breastfeeding among newly postpartum women with less than a high school education and increased anticipated breastfeeding when encouraged by the baby’s father and their own mother.

Family physicians should promote breastfeeding through provision of a positive office environment, anticipatory guidance, and ongoing communication with the patient and her social supports.^4,17 With active discussion and promotion of breastfeeding among urban, indigent communities, more pregnant women may consider, initiate, and maintain breastfeeding as the primary nutritional source for their infants.

Healthy People 2000, a national agenda for attention to health promotion and disease-prevention activities, has established a target that 75% of mothers will breastfeed at the time of their postpartum hospital discharge and 50% will continue to breastfeed through 6 months.¹ Breastfeeding offers several advantages, including the provision of a convenient balanced nutritional source, promotion of bonding between mother and infant, and conferral of an immunologic advantage because breastfed infants are less prone to respiratory and enteric infections.^2,3 Breastfeeding has also been reported to contribute to infant cognitive development and to decrease the risk of chronic diseases during childhood.⁴ Despite these documented benefits, the proportion of women who currently breastfeed their newborns remains suboptimal at 50%.¹

Breastfeeding practices vary by ethnicity, education, socioeconomic status, and maternal age.^5,6 The lowest rates are observed among young and undereducated mothers, as well as among African American mothers.^6,7 Among low-income groups, breastfeeding is initiated by approximately one third of mothers and is continued for 6 months by only 9%.⁷ Since low-income, minority populations commonly experience a higher prevalence of various infant health problems, breastfeeding seems especially important in benefiting infants from these communities.

Factors that influence breastfeeding have not been sufficiently studied among urban, low-income populations seen in family medicine centers. In an attempt to expand our knowledge of infant feeding practices among indigent populations, this research explored correlates of planned breastfeeding among postpartum women in a low-income, urban family practice setting.

Methods

This research project used a structured survey instrument to assess anticipated infant feeding practices among newly postpartum women from an urban family medicine site. This cross-sectional study was approved by the hospital Institutional Review Board. Informed consent was obtained from all participants.

Study Population

This study was completed among patients from an academic family medicine training site in Upstate New York. This facility-the Family Medicine Center (FMC), a community-based residency practice located in an urbanized area-offers a variety of medical services, including provision of comprehensive and longitudinal ambulatory family medical care. The FMC is situated in a health professional shortage area; a significant proportion of patients seen at the facility can be represented as economically disadvantaged and minority (eg, African American).

Study participants included postpartum women from the FMC who gave birth between January 1996 and June 1997. There were no restrictions based on race or age. Nearly 84% (67/80) of all eligible postpartum patients were invited to participate. No systematic patterns were identified among the subset of postpartum women (n=13) who did not receive survey forms, and no systematic bias was identified among the postpartum women not interviewed. Surveys were completed for 98.5% (66/67) of the patients asked to participate. Interviews were conducted while subjects were postpartum in-patients at an affiliated tertiary care hospital where patients from the FMC are routinely admitted for deliveries.

Survey Instrument

Using a structured survey instrument, we obtained data on demographics, relevant clinical history, and past infant feeding practices, as well as several items relating to current infant feeding plans and attitudes. Survey responses were based on both a medical chart review and an interview administered to subjects by trained individuals.

Data Analyses

Completed surveys were entered into a computerized database using SPSS/PC (SPSS/PC+, SPSS Inc., Chicago, Ill, 1990) and Epi Info (Epi Info version 5.01, Centers for Disease Control, Epidemiology Program Office, Atlanta, Ga, 1990). We used cross-tabulations and contingency table analyses to examine potential predictors of, as well as barriers to, breastfeeding. Independent variables included subject demographics, reproductive history, previous experiences with breastfeeding, and responses to attitudinal items on breastfeeding. We used odds ratios as one measure of association to assess the relationship between selected independent variables and anticipated infant breastfeeding practices. Ninety-five percent confidence intervals for the odds ratio provided an indication of both the precision and significance of these point estimates of risk. We used multiple logistic regression to calculate adjusted odds ratios and confidence intervals. Selected demographic, historical, and attitudinal variables found to be associated with breastfeeding through univariate analyses (P <.10) were entered into the logistic model.

Respondents were stratified into 2 groups on the basis of their responses to survey items on anticipated infant feeding practices. Subjects who reported plans to either exclusively breastfeed or plans to use any combination of breastfeeding and bottle-feeding were classified as the “breastfeed” group, while respondents planning to only bottle-feed their infants were classified as the “bottle-feed” group.

Results

Our results are based on interviews completed with 66 postpartum patients. The majority of participants were African American (95%) and single (95%), with a median age of 24 years. They completed a median of 12 years of schooling, had a median of 3 pregnancies, and a median 2 previous births. Selected demographic and reproductive characteristics are presented in Table 1.

Only 3 participants (4.5%) indicated that they planned to breastfeed exclusively, while an additional 11 (16.7%) noted plans to use a combination of bottle-feeding and breastfeeding. Breastfeeding plans were not found to be associated with maternal demographics features or with reproductive characteristics. A history of breastfeeding was reported by 28% of those planning to breastfeed their new infant, compared with 8% of those planning to bottle-feed; however, this difference failed to achieve significance (odds ratio [OR]=4.80; 95% confidence interval [CI], 0.74-29.7).

Analyses of attitudes about infant feeding are summarized in Table 2. Women planning to breastfeed their infants were more likely to agree that breastfeeding helped reduce their own weight after pregnancy and that they were encouraged to breastfeed by the baby’s father and by their mothers. Respondents planning to bottle-feed reported that breastfeeding was too complicated.

Multiple logistic regression yielded a model that accounted for 50% of the variance in anticipated breastfeeding practice. The adjusted model identified significant odds ratios for the association between low levels of maternal education and anticipated breastfeeding (OR=0.13; 95% CI, 0.05-0.35) and for encouragement from the baby’s father or the participant’s mother to breastfeed and anticipated breastfeeding (OR=12.4; 95% CI, 4.92-31.4). This logistic model also correctly classified 86.5% of the respondents, including 46% of the breastfeed group and 100% of the bottle-feed group. (Univariate odds ratios for these independent variables were 0.18 (0.02-0.95) for limited education and 8.60 (1.71-43.7) for encouragement.)

Discussion

The benefits of infant breastfeeding-including improved maternal-infant bonding, nutritional completeness and enhanced immunologic effects-are well documented and result in decreased infant morbidity during the first year of life.^2,3 This study contributes valuable data on anticipated infant feeding practices among predominantly African American patients from a low-income, urban community. While the goal of Healthy People 2000 is to attain a level of 75% of mothers who breastfeed in the immediate postpartum period, only 21% of study subjects indicated plans to breastfeed their infants. It should be noted that this figure likely represents an overestimate of actual infant feeding practices, since only 3 subjects (4.5%) reported plans to exclusively breastfeed their infants. Moreover, respondents planning to breastfeed noted intentions to nurse their infants for periods ranging from 4 to 52 weeks (median = 8 weeks) and for 30% to 100% of all feedings (median = 75% of feedings). The plans described by these mothers suggest a weak commitment to breastfeeding. For breastfeeding to be successful in the immediate postpartum period it is important that breast milk be the exclusive nutritional source. In addition, the plans for the limited duration of breastfeeding lag behind the public health goal to maintain a level of 50% breastfeeding at 6 months of age.

This study examined anticipated feeding, rather than actual feeding practices, to allow for data collection without the need for further follow-up after delivery. Previous research suggests that intention to breastfeed is related to both initiation^8,9 and duration¹⁰ of breastfeeding. Because, in part, of the limited numbers of study subjects planning to breastfeed their infants (n=14), many of our comparisons failed to achieve statistical significance. Using the chi-square test of independence, the observed sample sizes, and a 2-sided test with a level of significance of 5%, there was approximately 80% power to detect absolute differences of 30% to 35% in the self-reported infant feeding plans by demographic group or attitudinal response. This study design might be replicated among a larger number of postpartum patients at other family medicine centers.

Several significant associations were identified, including limited maternal educational attainment and decreased likelihood of infant breastfeeding (OR=0.18). This association between breastfeeding and education has been noted previously.^5,6 In addition, encouragement from the subject’s mother was noted to increase the likelihood of breastfeeding 12-fold. The broad application of breastfeeding promotion programs to include members of a woman’s social support system has been suggested as a means of increasing breastfeeding rates.^9,11,12 Similarly, the lack of social support for breastfeeding has been associated with increased bottle-feeding.¹³

A recent article reported limited knowledge of breastfeeding issues and techniques among family medicine attending physicians and residents.¹⁴ To address these knowledge gaps findings from this study should be used to develop educational interventions that actively promote breastfeeding among members of this community and in similar patient populations. Family medicine centers, for example, could develop a family-centered intervention whereby mothers and grandmothers of pregnant patients might use a structured forum to encourage breastfeeding by reviewing known benefits and sharing stories of successful breastfeeding experiences. The importance of soliciting and maintaining the involvement of the baby’s father in the promotion of breastfeeding is supported by our study, as well as by previous reports.^8,15-17 In this educational context, fathers might be persuaded to attend prenatal visits where the topic of infant feeding is addressed. Various financial incentives could also be used to stimulate the participation of urban, minority pregnant women and their families and partners in educational workshops.¹⁸ These forums should provide opportunities to address the prevalent misperceptions regarding the effects and benefits of breastfeeding.

Conclusions

This study presents unique data regarding anticipated infant feeding practices among predominantly African American patients from a low-income, urban community. The results reveal decreased anticipated breastfeeding among newly postpartum women with less than a high school education and increased anticipated breastfeeding when encouraged by the baby’s father and their own mother.

Family physicians should promote breastfeeding through provision of a positive office environment, anticipatory guidance, and ongoing communication with the patient and her social supports.^4,17 With active discussion and promotion of breastfeeding among urban, indigent communities, more pregnant women may consider, initiate, and maintain breastfeeding as the primary nutritional source for their infants.

Healthy People 2000, a national agenda for attention to health promotion and disease-prevention activities, has established a target that 75% of mothers will breastfeed at the time of their postpartum hospital discharge and 50% will continue to breastfeed through 6 months.¹ Breastfeeding offers several advantages, including the provision of a convenient balanced nutritional source, promotion of bonding between mother and infant, and conferral of an immunologic advantage because breastfed infants are less prone to respiratory and enteric infections.^2,3 Breastfeeding has also been reported to contribute to infant cognitive development and to decrease the risk of chronic diseases during childhood.⁴ Despite these documented benefits, the proportion of women who currently breastfeed their newborns remains suboptimal at 50%.¹

Breastfeeding practices vary by ethnicity, education, socioeconomic status, and maternal age.^5,6 The lowest rates are observed among young and undereducated mothers, as well as among African American mothers.^6,7 Among low-income groups, breastfeeding is initiated by approximately one third of mothers and is continued for 6 months by only 9%.⁷ Since low-income, minority populations commonly experience a higher prevalence of various infant health problems, breastfeeding seems especially important in benefiting infants from these communities.

Factors that influence breastfeeding have not been sufficiently studied among urban, low-income populations seen in family medicine centers. In an attempt to expand our knowledge of infant feeding practices among indigent populations, this research explored correlates of planned breastfeeding among postpartum women in a low-income, urban family practice setting.

Methods

This research project used a structured survey instrument to assess anticipated infant feeding practices among newly postpartum women from an urban family medicine site. This cross-sectional study was approved by the hospital Institutional Review Board. Informed consent was obtained from all participants.

Study Population

This study was completed among patients from an academic family medicine training site in Upstate New York. This facility-the Family Medicine Center (FMC), a community-based residency practice located in an urbanized area-offers a variety of medical services, including provision of comprehensive and longitudinal ambulatory family medical care. The FMC is situated in a health professional shortage area; a significant proportion of patients seen at the facility can be represented as economically disadvantaged and minority (eg, African American).

Study participants included postpartum women from the FMC who gave birth between January 1996 and June 1997. There were no restrictions based on race or age. Nearly 84% (67/80) of all eligible postpartum patients were invited to participate. No systematic patterns were identified among the subset of postpartum women (n=13) who did not receive survey forms, and no systematic bias was identified among the postpartum women not interviewed. Surveys were completed for 98.5% (66/67) of the patients asked to participate. Interviews were conducted while subjects were postpartum in-patients at an affiliated tertiary care hospital where patients from the FMC are routinely admitted for deliveries.

Survey Instrument

Using a structured survey instrument, we obtained data on demographics, relevant clinical history, and past infant feeding practices, as well as several items relating to current infant feeding plans and attitudes. Survey responses were based on both a medical chart review and an interview administered to subjects by trained individuals.

Data Analyses

Completed surveys were entered into a computerized database using SPSS/PC (SPSS/PC+, SPSS Inc., Chicago, Ill, 1990) and Epi Info (Epi Info version 5.01, Centers for Disease Control, Epidemiology Program Office, Atlanta, Ga, 1990). We used cross-tabulations and contingency table analyses to examine potential predictors of, as well as barriers to, breastfeeding. Independent variables included subject demographics, reproductive history, previous experiences with breastfeeding, and responses to attitudinal items on breastfeeding. We used odds ratios as one measure of association to assess the relationship between selected independent variables and anticipated infant breastfeeding practices. Ninety-five percent confidence intervals for the odds ratio provided an indication of both the precision and significance of these point estimates of risk. We used multiple logistic regression to calculate adjusted odds ratios and confidence intervals. Selected demographic, historical, and attitudinal variables found to be associated with breastfeeding through univariate analyses (P <.10) were entered into the logistic model.

Respondents were stratified into 2 groups on the basis of their responses to survey items on anticipated infant feeding practices. Subjects who reported plans to either exclusively breastfeed or plans to use any combination of breastfeeding and bottle-feeding were classified as the “breastfeed” group, while respondents planning to only bottle-feed their infants were classified as the “bottle-feed” group.

Results

Our results are based on interviews completed with 66 postpartum patients. The majority of participants were African American (95%) and single (95%), with a median age of 24 years. They completed a median of 12 years of schooling, had a median of 3 pregnancies, and a median 2 previous births. Selected demographic and reproductive characteristics are presented in Table 1.

Only 3 participants (4.5%) indicated that they planned to breastfeed exclusively, while an additional 11 (16.7%) noted plans to use a combination of bottle-feeding and breastfeeding. Breastfeeding plans were not found to be associated with maternal demographics features or with reproductive characteristics. A history of breastfeeding was reported by 28% of those planning to breastfeed their new infant, compared with 8% of those planning to bottle-feed; however, this difference failed to achieve significance (odds ratio [OR]=4.80; 95% confidence interval [CI], 0.74-29.7).

Analyses of attitudes about infant feeding are summarized in Table 2. Women planning to breastfeed their infants were more likely to agree that breastfeeding helped reduce their own weight after pregnancy and that they were encouraged to breastfeed by the baby’s father and by their mothers. Respondents planning to bottle-feed reported that breastfeeding was too complicated.

Multiple logistic regression yielded a model that accounted for 50% of the variance in anticipated breastfeeding practice. The adjusted model identified significant odds ratios for the association between low levels of maternal education and anticipated breastfeeding (OR=0.13; 95% CI, 0.05-0.35) and for encouragement from the baby’s father or the participant’s mother to breastfeed and anticipated breastfeeding (OR=12.4; 95% CI, 4.92-31.4). This logistic model also correctly classified 86.5% of the respondents, including 46% of the breastfeed group and 100% of the bottle-feed group. (Univariate odds ratios for these independent variables were 0.18 (0.02-0.95) for limited education and 8.60 (1.71-43.7) for encouragement.)

Discussion

The benefits of infant breastfeeding-including improved maternal-infant bonding, nutritional completeness and enhanced immunologic effects-are well documented and result in decreased infant morbidity during the first year of life.^2,3 This study contributes valuable data on anticipated infant feeding practices among predominantly African American patients from a low-income, urban community. While the goal of Healthy People 2000 is to attain a level of 75% of mothers who breastfeed in the immediate postpartum period, only 21% of study subjects indicated plans to breastfeed their infants. It should be noted that this figure likely represents an overestimate of actual infant feeding practices, since only 3 subjects (4.5%) reported plans to exclusively breastfeed their infants. Moreover, respondents planning to breastfeed noted intentions to nurse their infants for periods ranging from 4 to 52 weeks (median = 8 weeks) and for 30% to 100% of all feedings (median = 75% of feedings). The plans described by these mothers suggest a weak commitment to breastfeeding. For breastfeeding to be successful in the immediate postpartum period it is important that breast milk be the exclusive nutritional source. In addition, the plans for the limited duration of breastfeeding lag behind the public health goal to maintain a level of 50% breastfeeding at 6 months of age.

This study examined anticipated feeding, rather than actual feeding practices, to allow for data collection without the need for further follow-up after delivery. Previous research suggests that intention to breastfeed is related to both initiation^8,9 and duration¹⁰ of breastfeeding. Because, in part, of the limited numbers of study subjects planning to breastfeed their infants (n=14), many of our comparisons failed to achieve statistical significance. Using the chi-square test of independence, the observed sample sizes, and a 2-sided test with a level of significance of 5%, there was approximately 80% power to detect absolute differences of 30% to 35% in the self-reported infant feeding plans by demographic group or attitudinal response. This study design might be replicated among a larger number of postpartum patients at other family medicine centers.

Several significant associations were identified, including limited maternal educational attainment and decreased likelihood of infant breastfeeding (OR=0.18). This association between breastfeeding and education has been noted previously.^5,6 In addition, encouragement from the subject’s mother was noted to increase the likelihood of breastfeeding 12-fold. The broad application of breastfeeding promotion programs to include members of a woman’s social support system has been suggested as a means of increasing breastfeeding rates.^9,11,12 Similarly, the lack of social support for breastfeeding has been associated with increased bottle-feeding.¹³

A recent article reported limited knowledge of breastfeeding issues and techniques among family medicine attending physicians and residents.¹⁴ To address these knowledge gaps findings from this study should be used to develop educational interventions that actively promote breastfeeding among members of this community and in similar patient populations. Family medicine centers, for example, could develop a family-centered intervention whereby mothers and grandmothers of pregnant patients might use a structured forum to encourage breastfeeding by reviewing known benefits and sharing stories of successful breastfeeding experiences. The importance of soliciting and maintaining the involvement of the baby’s father in the promotion of breastfeeding is supported by our study, as well as by previous reports.^8,15-17 In this educational context, fathers might be persuaded to attend prenatal visits where the topic of infant feeding is addressed. Various financial incentives could also be used to stimulate the participation of urban, minority pregnant women and their families and partners in educational workshops.¹⁸ These forums should provide opportunities to address the prevalent misperceptions regarding the effects and benefits of breastfeeding.

Conclusions

This study presents unique data regarding anticipated infant feeding practices among predominantly African American patients from a low-income, urban community. The results reveal decreased anticipated breastfeeding among newly postpartum women with less than a high school education and increased anticipated breastfeeding when encouraged by the baby’s father and their own mother.

Family physicians should promote breastfeeding through provision of a positive office environment, anticipatory guidance, and ongoing communication with the patient and her social supports.^4,17 With active discussion and promotion of breastfeeding among urban, indigent communities, more pregnant women may consider, initiate, and maintain breastfeeding as the primary nutritional source for their infants.