Improved Estimation of Breastfeeding Rates Using a Novel Breastfeeding and Milk Expression Survey

Abstract

Background:

Increasing the proportion of infants who are breastfed and extending breastfeeding duration are high-priority U.S. goals. Evaluation of progress is based on federal survey data, but federal survey questions do not reflect contemporary feeding practices.

Materials and Methods:

Our objective was to evaluate the Brief Breastfeeding and Milk Expression Recall Survey (BaByMERS) in estimating breast milk feeding and milk expression practices and compare to estimates from simultaneously administered federal survey questions. We surveyed women with child(ren) younger than the age of 6 years attending a large children's hospital for urgent or primary care. We estimated the proportions who participated in various breast milk feeding and milk expression practices and the durations of each and examined agreement between the surveys. We compared respondents with high versus low disagreement using log-binomial regression.

Results:

Of 225 respondents, 51% had less than a Bachelor's degree, and 44% identified as a race other than white. Similar proportions on each survey reported ever having breastfed or fed breast milk (84%). Proportions still breastfeeding or feeding breast milk at 6 and 12 months differed slightly by survey. Dyads (9%) who fed at the breast and fed expressed milk for nonidentical periods had estimates for the duration of breastfeeding or feeding breast milk that were lower per the federal survey. Respondents who answered the federal survey before the BaByMERS were more likely to provide discrepant responses (risk ratio = 3.40, 95% confidence interval: 1.18–9.80).

Conclusions:

This study offers further validation of brief interviewer-administered questions to collect quality data recalled about infant feeding and lactation for research purposes.

Introduction

Increasing the proportion of infants who are breastfed and extending breastfeeding duration are high-priority public health goals for the United Sates. Evaluation of progress is based on national survey data, specifically the National Immunization Survey (NIS).¹ Based on the NIS, an estimated 83% of U.S. infants born in 2015 were ever breastfed, and only 58% were still breastfeeding at 6 months of age.² A recalled infant feeding history serves as a key early childhood health indicator in numerous other major national surveys such as the National Health and Nutrition Examination Survey (NHANES) and the National Survey of Children's Health (NSCH).^3,4

Many federal surveys, including the NIS, the NHANES, and the NSCH, use the following questions: “Was [child] ever breastfed or fed breast milk?” And if so, “How old was [child] when [he/she] completely stopped breastfeeding or being fed breast milk?” These questions are used to calculate two important indicators: 1) the proportion of U.S. infants who are ever breastfed or fed breast milk and 2) the duration of breastfeeding or feeding breast milk for U.S. children. Although infant feeding practices have changed dramatically recently, the questions have not evolved to accommodate these changes. For instance, as of 2006, at least 85% of women who provided breast milk for their infant did so using a pump at least sometimes, compared to only 38% in 1993.^5,6 Five to seven percent of breast milk fed infants are never fed at the breast.^6,7 Some families store expressed milk for their child to consume contemporaneously with feeding at the breast or for after lactation has ceased, while some women continue to lactate after their child has been weaned and discard or donate the milk.⁸ It is unknown whether respondents fully consider expressed milk feeding in their responses to the national surveys, or if they think only or mostly of feeding at the breast when they respond. If respondents do not completely report both expressed milk feeding and feeding at the breast, national statistics for the duration of breastfeeding would be underestimated. Infants who are fed exclusively expressed or donor milk may be counted as having never been breastfed or fed breast milk at all if respondents only consider feeding at the breast in their response. For research about maternal lactation, only the National Survey of Family Growth has relevant questions and those who do not distinguish feeding at the breast from milk expression.⁹

Recent evidence suggests that feeding breast milk from a bottle has attenuated benefits compared to feeding at the breast in terms of respiratory, growth, and gastrointestinal outcomes.^10–13 As a result, it is important that expressed milk feeding be estimated distinctly from feeding at the breast, but the existing federal survey questions cannot do this.

There is, therefore, a need to develop new, succinct questions about breast milk feeding practices that can improve data quality and be incorporated efficiently into both federal and nonfederal surveys that require recall of breast milk feeding practices. We recently reported results from initial validation of the Brief Breastfeeding and Milk Expression Recall Survey (BaByMERS) based on cognitive interviews.¹⁴ The next step, and objective of the present study, was to administer the BaByMERS alongside the federal survey questions used in the NIS, NHANES, and NSCH to examine agreement between the two surveys in estimating breast milk feeding and maternal milk expression practices.

Materials and Methods

Surveys

We adapted the questions in the BaByMERS from our prior Moms2Moms Study where mothers recalled at 12 months postpartum their child's age when the child started and stopped feeding at the breast and started and stopped consuming expressed milk, and when the mother started and stopped expressing milk.¹² The original Moms2Moms questions were based on the suggestions of Geraghty and Rasmussen.¹⁵ The current version of BaByMERS and the federal survey questions are presented in Table 1.

Table 1.

Brief Breastfeeding and Milk Expression Recall Survey and Infant Feeding Questions from U.S. Federal Surveys (n = 225, 2017–18, Ohio)

Construct measured	BaByMERS		U.S. Federal Surveys
Construct measured	Questions	Answer options	Questions	Answer options
Feeding at the breast	B1. Was [child] ever breastfed directly at the breast?	Yes \| No \| Refused \| Don't know [if No, Refused, Don't know, skip to next Instruction]
	B2. How old was [child] when [he/she] started feeding at the breast?	Number in days, weeks, or months \| Refused \| Don't know
	B3. How old was [child] when [he/she] completely stopped feeding at the breast?	Number in days, weeks, or months \| Still feeding at the breast \| Refused \| Don't know
Maternal milk expression	The next couple of questions are about expressing your breast milk using a breast pump or your hands.
	B4. Did you ever use a breast pump or your hands to express milk for [child]?	Yes \| No \| Refused \| Don't know [if No, Refused, Don't know, skip to B7]
	B5. How old was [child] when you started using a breast pump or your hands to express your breast milk?	Number in days, weeks, or months \| Refused \| Don't know
	B6. How old was [child] when you stopped using a breast pump or your hands to express your breast milk?	Number in days, weeks, or months \| Still expressing breast milk \| Refused \| Don't know
Child's consumption of expressed milk	B7. Did [child] ever drink expressed breast milk?	Yes \| No \| Refused \| Don't know [if No, Refused, Don't know, stop]
	B8. How old was [child] when [he/she] started drinking expressed breast milk?	Number in days, weeks, or months \| Refused \| Don't know
	B9. How old was [child] when [he/she] completely stopped drinking expressed breast milk? Include fresh and frozen breast milk and breast milk mixed in cereal.	Number in days, weeks, or months \| Still drinking expressed breast milk \| Refused \| Don't know
Child ever breastfed or fed breast milk	Derived from B1 and B7 above	If B1 = “Yes” or B7 = “Yes” then Yes. If B1 = “No” and B7 = “No” then No.	F1. Was [child] ever breastfed or fed breast milk?	Yes \| No \| Refused \| Don't know [if No, Refused, Don't know, stop]
Duration child was breastfed or fed breast milk	Derived from B2, B3, B8, B9 above	Maximum value among (B3, B9) minus minimum value among (B2, B8)	F2. How old was [child] when [he/she] completely stopped breastfeeding or being fed breast milk?	Number in days, weeks, or months \| Still breastfeeding \| Refused \| Don't know

BaByMERS, Brief Breastfeeding and Milk Expression Recall Survey.

Participant recruitment

In 2017–2018, we approached women waiting with their child(ren) in Urgent and Primary Care Clinics at Nationwide Children's Hospital, Columbus, OH, USA and invited their participation in an infant feeding interview. Eligible women were English-speaking, 18 years or older, and the biologic mother of at least one child 12 months to 5 years old. The age range for children was selected to mirror the eligibility for the NSCH module about breastfeeding.³ The interviewer administered an informed consent script, and consent was electronically documented. This study was reviewed and approved by the Nationwide Children's Hospital Institutional Review Board.

Interview procedures

A trained interviewer asked how many biologic, singleton children the woman had, and a computer algorithm programmed in the electronic survey research tool, REDCap, randomly selected one child younger than the age of 6 years to be the subject of the subsequent feeding questions.¹⁶ The interview proceeded first to the BaByMERS and then the federal survey questions or vice versa, per a concealed randomization scheme in REDCap. This was to permit evaluation of ordering effects. The interview concluded with a demographics module, and the interviewer entered data into REDCap in real-time using a laptop computer. REDCap time stamped each section to enable calculation of completion times. Respondents received a parking pass or $5 value incentive.

Study variables

We used the BaByMERS questions with yes/no response options (B1, B4, and B7) to calculate the proportions of dyads who ever/never fed at the breast, women who ever/never expressed milk, and children who were ever/never fed expressed milk. We considered those who marked “yes” either to the question about having ever fed at the breast (B1) or the question about having ever fed their child expressed milk (B7) as having a child who was ever breastfed or fed breast milk. We considered those who marked “no” to both of those questions to have a child who was never breastfed or fed breast milk. We combined B1 and B7 in that way to mimic the federal survey question F1. We calculated duration variables for each of these feeding practices by subtracting the child's age when they (or the woman) started participating in that particular feeding practice from the child's age when they stopped, in days.

The federal survey questions enabled calculation of the proportion of respondents whose child was ever/never breastfed or fed breast milk (F1) and the duration of this feeding practice based on the child's age when they stopped (the start date was not asked) (F2), but it did not enable estimation of the other feeding practices that BaByMERS could estimate. Thus, it was possible to directly compare the two surveys only on their estimates of ever/never and duration of breastfeeding or feeding breast milk.

We calculated the proportion of children still breastfeeding or being fed breast milk at 6 and 12 months of age based on each survey. These time points mirror national benchmarks as reported by the Centers for Disease Control and Prevention (CDC).²

Statistical analysis

Using univariate statistics, we characterized the sample and estimated proportions, medians, and interquartile ranges for each feeding practice for each survey. For duration variables, we set to missing any respondents who did not participate in that feeding practice. We evaluated agreement between the two surveys for the feeding practices that were captured by both: whether the child had ever been breastfed or fed breast milk (κ statistic), whether the child was still being breastfed or fed breast milk at 6 and 12 months of age (κ statistic), and the duration of breastfeeding or feeding breast milk (intraclass correlation coefficient [ICC], Bland–Altman plot). We examined ordering effects by calculating κ and ICC separately for each randomized group. We used log-binomial regression to compare respondents who had high disagreement (>14 days) between the surveys in their values for the duration of breastfeeding or feeding breast milk to those with low disagreement (≤14 days), by which survey they completed first sociodemographics.

Results

The sample of 225 women was diverse in age, education (about one-quarter had a high school diploma or less education and one-quarter had a graduate or professional degree), race and ethnicity (22% were black or African American, 5% were Hispanic), and income (Table 2).

Table 2.

Participant Characteristics (n = 225, 2017–18, Ohio)

Maternal or child characteristic	Total sample, n (%)
Maternal age
18–27	53 (24)
28–31	47 (21)
32–35	55 (24)
≥36	69 (31)
Missing	1 (0)
Maternal education
Less than high school diploma	21 (9)
High school graduate or GED or equivalent	41 (18)
Some college or Associate degree	53 (24)
Bachelor's degree	57 (25)
Graduate or professional degree	52 (23)
Missing	1 (0)
Marital status
Married	149 (66)
Living with partner	26 (12)
Separated, divorced, widowed, or never married	49 (22)
Missing	1 (0)
Maternal race
White	149 (66)
Black or African American	50 (22)
Asian	10 (4)
Other race	8 (4)
Multiple races	7 (3)
Missing	1 (0)
Maternal ethnicity
Hispanic	12 (5)
Non-Hispanic	213 (95)
Household income
<$15,000	36 (16)
$15,000–< $40,000	44 (20)
$40,000–< $70,000	27 (12)
≥$70,000	109 (48)
Missing	9 (4)
Number of biologic children living with mother
1	146 (65)
2	67 (30)
≥3	12 (5)
Child sex
Female	115 (51)
Male	106 (47)
Missing	4 (2)
Child age at time of survey
1 year	54 (24)
2 years	39 (17)
3 years	44 (20)
4 years	55 (24)
5 years	33 (15)

GED, general educational development.

The BaByMERS was quick to administer, requiring a mean of 52 seconds to complete (SD = 36). The diverse ways that infants were fed breast milk and women expressed milk, as captured by the BaByMERS, is displayed in Figure 1. Based on the BaByMERS, 79% reported having ever fed at the breast, 79% reported having ever expressed milk, 69% reported their child ever consumed expressed milk, and 84% reported having ever breastfed or fed their child breast milk (Table 3). The federal survey could not estimate the proportions that ever fed at the breast, ever expressed milk, or ever consumed expressed milk, just the proportion that was ever breastfed or fed breast milk (83%). Three women reported on the BaByMERS that they expressed milk, but their child was never fed at the breast and never consumed expressed milk. The federal survey accurately classified their children as having never fed at the breast or fed breast milk.

FIG. 1.

Proportions of mothers and children who ever participated in the three breast milk feeding and milk expression practices measured using the BaByMERS. Thirty-four (15%) children were exclusively formula fed and are not shown. Seven women (3%) responded “Don't Know” or declined or were unable to answer whether their child was fed expressed milk (all of them fed at the breast and 6 of them expressed milk), and they are not shown. BaByMERS, Brief Breastfeeding and Milk Expression Recall Survey.

Table 3.

Breast Milk Feeding and Milk Expression Practices as Reported Using the Brief Breastfeeding and Milk Expression Recall Survey and the U.S. Federal Surveys (n = 225, 2017–18, Ohio)

	BaByMERS	Federal survey
Feeding at the breast		—
Child was ever fed at the breast, n (%)	177 (79)
Child's age when started, median (IQR), days	0 (0)	—
Child's age when stopped, median (IQR), days	180 (300)	—
Duration of feeding at the breast,^a median (IQR), days	180 (300)	—
Maternal milk expression		—
Mother ever expressed milk, n (%)	178 (79)
Child's age when started, median (IQR), days	6 (33)	—
Child's age when stopped, median (IQR), days	180 (216)	—
Duration of maternal milk expression,^a median (IQR), days)	150 (199)	—
Child's consumption of expressed milk		—
Child ever consumed expressed milk, n (%)	156 (69)
Child's age when started, median (IQR), days	13 (53)	—
Child's age when stopped, median (IQR), days	180 (270)	—
Duration of consuming expressed milk,^a median (IQR), days	179 days (234)	—
Child ever breastfed or fed breast milk, n (%)	188 (84)	187 (83)
Child was ever breastfed or fed breast milk, n (%)	188 (84)
Child's age when started, median (IQR), days	0 (0)	—
Child's age when stopped, median (IQR), days	180 (270)	180 (270)
Duration of breastfeeding or feeding breast milk,^a median (IQR), days	180 (270)	180 (270)^b
Breastfeeding or feeding breast milk at 6 months, n (%)	92 (41%)	86 (38%)
Breastfeeding or feeding breast milk at 12 months, n (%)	42 (19%)	40 (18%)

Durations are calculated among those who ever participated in the feeding practice (i.e., nonzero duration values only). Respondents who answered No or Don't Know or refused to answer the question about whether they had ever participated in a particular practice (n = 7 for child consumption of expressed milk, 0 for the other practices) or who were unable to provide a value for the child's age when they started or stopped the practice (n = 15) were considered to have a missing value for the duration of that practice. If duration was missing for feeding at the breast or child consumption of expressed milk but not both, the duration the child was ever breastfed or fed breast milk was taken to be the value for the nonmissing practice.

To calculate duration, the federal surveys assume that breastfeeding or feeding breast milk begins on the first day of life.

IQR, interquartile range.

The median duration of breastfeeding or feeding breast milk was 180 days (interquartile range = 270) based on either survey. The proportions of the sample still breastfeeding or feeding breast milk at 6 and 12 months differed by survey. Based on BaByMERS, it was 41% at 6 months and 19% at 12 months, while the federal survey estimated 38% and 18%, respectively. Estimated durations of feeding at the breast, maternal milk expression, and child consumption of expressed milk based on the BaByMERS are shown in Table 3. The federal survey could not estimate these durations.

Agreement between the BaByMERS and the federal survey

We observed very high agreement between the two surveys for whether the child was ever breastfed or fed breast milk (κ = 0.98; 95% confidence interval [CI]: 0.95–1.00). Only one woman had inconsistent responses across the two surveys on this item (“yes” on the BaByMERS, “no” on the federal survey). The two surveys also agreed a high proportion of the time on whether a child was still breastfeeding or being fed breast milk at 6 months (κ = 0.95; 95% CI: 0.91–0.99) and 12 months (κ = 0.95; 95% CI: 0.90–1.00). Agreement was also very high between the two surveys on the duration of breastfeeding or feeding breast milk based on the ICC (ICC = 0.99). However, the Bland–Altman plot showed some disagreement for that duration measure: the mean difference was 5.8 days (SD = 21.3) indicating systematic bias toward the BaByMERS generating higher values for duration than the federal survey (Fig. 2). The 95% limits of agreement were moderately wide, and seven observations (5%) fell outside the limits.

FIG. 2.

Bland–Altman Plot displaying the difference in breastfeeding or breast milk feeding duration (value from BaByMERS minus value from U.S. federal surveys) versus the mean of the two measures (n = 136, 2017–18, Ohio). Children (n = 89) who were exclusively formula fed, who were still breastfeeding or consuming breast milk, or who were with missing data required to calculate duration were excluded.

When the two surveys produced different values for the duration of breastfeeding or feeding breast milk of more than a few days, it was almost always for dyads who reported on the BaByMERS that their child continued consuming expressed milk after the dyad had stopped feeding at the breast, or vice versa. When asked the federal survey questions, these respondents omitted entirely or underreported the length of time the dyad continued with one practice after stopping the other, thereby producing the disagreement. Of the 79 dyads who reported on the BaByMERS that their child continued consuming expressed milk after the dyad had stopped feeding at the breast, or vice versa, 20 had a lower value for the duration of breastfeeding or breast milk feeding on the federal survey than on the BaByMERS, while 59 had identical values.

Examination of ordering effects revealed no meaningful differences between the randomized groups in agreement between the two surveys for whether the respondent ever breastfed or fed their child breast milk (κ = 1.0 for the group that received the BaByMERS first, κ = 0.96 for the group that received the federal survey first) and for the total duration of breastfeeding or feeding breast milk (ICC = 0.99 for both groups). However, while most women had low disagreement (≤14 days), ∼13% of the sample had high disagreement (>14 days) between the surveys in the duration of breastfeeding or feeding breast milk. Respondents who completed the federal survey first were three times more likely to have high disagreement compared to respondents who completed the BaByMERS first (risk ratio [RR] = 3.40, 95% CI: 1.18–9.80). In fact, of those who completed the federal survey first, 20% had high disagreement, while only 6% of those who completed the BaByMERS first had high disagreement. Among the sociodemographics, only Hispanic ethnicity was associated with an increased risk of high disagreement between the surveys in reported duration of breastfeeding or feeding breast milk compared to non-Hispanic women (RR = 3.69, 95% CI: 1.38–9.81) (Table 4). Other effect estimates were elevated but their CIs were imprecise (e.g., race).

Table 4.

Disagreement in Reported Duration of Breastfeeding or Feeding Breast milk Between the Brief Breastfeeding and Milk Expression Recall Survey and the U.S. Federal Surveys in Relationship to Respondent Characteristics and Questionnaire Ordering (n = 136, 2017–18, Ohio)

	>14 days' disagreement n = 18, n (%)	≤14 days' disagreement n = 118, n (%)	Risk ratio (95% confidence interval)
Maternal age
18–27	4 (22)	17 (15)	Reference
28–31	3 (17)	22 (19)	0.63 (0.16–2.50)
32–35	5 (28)	34 (29)	0.67 (0.20–2.24)
≥36	6 (33)	44 (38)	0.63 (0.20–2.01)
Maternal education
Less than high school diploma	1 (6)	4 (3)	Reference
High school graduate or GED or equivalent	5 (28)	13 (11)	1.39 (0.21–9.33)
Some college or Associate degree	4 (22)	24 (21)	0.71 (0.10–5.14)
Bachelor's degree	4 (22)	38 (32)	0.48 (0.07–3.47)
Graduate or professional degree	4 (22)	38 (32)	0.48 (0.07–3.47)
Marital status
Married	14 (78)	88 (75)	Reference
Living with partner	2 (11)	10 (9)	1.21 (0.31–4.71)
Separated, divorced, widowed, or never married	2 (11)	19 (16)	0.69 (0.17–2.83)
Maternal race
White	17 (94)	86 (74)	5.28 (0.73–38.15)
Black or African American, Asian, other, or multiple races	1 (6)	31 (27)	Reference
Maternal ethnicity
Hispanic	3 (17)	4 (3)	3.69 (1.38–9.81)
Non-Hispanic	15 (83)	114 (97)	Reference
Household income
<$15,000	2 (12)	13 (11)	Reference
$15,000–< $40,000	3 (18)	15 (13)	1.25 (0.24–6.53)
$40,000–< $70,000	3 (18)	14 (12)	1.32 (0.25–6.88)
≥$70,000	9 (53)	72 (63)	0.83 (0.20–3.48)
Number of biologic children living with mother
1	14 (78)	78 (66)	Reference
2	2 (11)	36 (31)	0.35 0.08–1.45)
≥3	2 (11)	4 (3)	2.19 (0.64–7.49)
Child sex
Female	9 (53)	60 (51)	1.06 (0.44–2.58)
Male	8 (47)	57 (49)	Reference
Child age at time of survey
1 year	5 (28)	27 (23)	Reference
2 years	3 (17)	23 (19)	0.74 (0.19–2.80)
3 years	3 (17)	24 (20)	0.71 (0.19–2.71)
4 years	5 (28)	28 (24)	0.97 (0.31–3.03)
5 years	2 (11)	16 (14)	0.71 (0.15–3.30)
Questionnaire ordering
Respondent completed the federal survey first	14 (78)	55 (47)	3.40 (1.18–9.80)
Respondent completed the BaByMERS first	4 (22)	63 (53)	Reference

Discussion

This study to validate the BaByMERS as a brief tool to capture a recalled history of feeding at the breast, maternal milk expression, and child consumption of expressed milk with parents of young children confirmed that the BaByMERS offers several advantages. The BaByMERS distinctly estimates feeding at the breast, maternal milk expression, and child consumption of expressed milk, while the federal survey estimates only a composite of breastfeeding and feeding breast milk. For the two measures that both surveys estimate (ever/never breastfeeding or feeding breast milk, and its duration), the surveys performed very similarly. The proportion still breastfeeding or being fed breast milk at 6 months was perhaps slightly higher for the BaByMERS, but still very similar. Based on calculation of medians for the whole sample, the two surveys produced the same results for the duration of breastfeeding or feeding breast milk. However, differences at the individual level were observed. For instance, respondents whose child continued consuming expressed breast milk after the time the dyad had stopped feeding at the breast, or the dyad had continued feeding at the breast after the child had stopped consuming expressed milk appeared to underreport the age of the child when he/she stopped breastfeeding or being fed breast milk on the federal survey. This affected 11% (20 of 176) of the respondents who had values from both surveys for the duration of breastfeeding or feeding breast milk. By asking separately about feeding at the breast, maternal milk expression, and child consumption of expressed milk, the BaByMERS may improve recall about these practices, thereby resulting in higher values based on the BaByMERS for overall duration of breastfeeding or feeding breast milk. This conclusion is supported by the fact that respondents who completed the federal survey first were more than three times more likely to have high disagreement across the two surveys.

The particular federal survey used as a comparison in this study was not designed to characterize maternal lactation. Women who expressed milk but their child was never fed at the breast and never fed expressed milk would be classified as having never breastfed or fed breast milk to their child if researchers use that federal survey to study maternal lactation history. The NSFG includes questions to capture a maternal lactation history, but they do not distinguish milk expression from feeding at the breast. The BaByMERS is designed to overcome these limitations.

National statistics based on the duration of breastfeeding or feeding breast milk, like the proportions of infants still breastfeeding or being fed breast milk at 6 and 12 months, are commonly referenced and used to track progress toward the Healthy People 2020 goals “Increase the proportion of infants who are breastfed at 6 months” and “Increase the proportion of infants who are breastfed at 12 months.”¹⁷ In fact, while the proportion of infants who have ever been breastfed or fed breast milk has increased recently to record highs, the United States continues to lag other countries in breastfeeding continuation.¹⁸ While the differences between the two surveys in the proportion of infants still breastfeeding or being fed breast milk at 6 and 12 months was only 1–2% points in this study, those may be meaningful differences when considering that the latest CDC data indicate that the United Sates is only 3% points away from meeting the Healthy People 2020 goal for the proportion of infants still breastfeeding at 6 months of age.²

In addition to considerations about national statistics, an urgent need exists to better understand the impact of milk expression and expressed milk feeding on maternal and child health because these practices are now extremely common, as is clear in this sample as well as others.^6,7 Several studies have linked expressed milk feeding to poorer child health outcomes, including coughing/wheezing, asthma, and altered growth.^10,13,19,20However, only a handful of studies have included questions to at least partially disentangle feeding mode from the substance being fed, including the Infant Feeding Practices Study II (IFPS II), the Canadian Healthy Infant Longitudinal Development study (CHILD), and our Moms2Moms study^6,12,13 The IFPS II questions and the published FeedCat Tool distinguish feeding modes, but were designed for repeated prospective administration that may not be cost-effective for all studies or for certain study designs.^21,22 The FeedCat Tool is also resource-intensive because it requires coding open-ended responses. The Questionnaire on Infant Feeding was designed for retrospective studies, but its length and self-administered format may not be a fit for interviewer-administered surveys.²³ Thus, the BaByMERS may fill a gap in the tools available for recalling infant feeding in large studies where truly prospective measurement may not be possible, and it requires less than 1 minute to complete.

This study was limited in that the sample was not drawn using complex sampling procedures to be nationally representative. As a result, the estimates cannot necessarily be generalized to the U.S. population. The limited sample size also precluded more precise effect estimates. Another limitation is that the length of the BaByMERS may preclude its adoption in very large studies and studies for which breastfeeding is a minor theme. Researchers uninterested in maternal lactation practices could omit questions B4–B6 to shorten it. The BaByMERS does not capture formula feeding, so studies interested in formula or breastfeeding exclusivity would require additional questions. Such questions are already available.¹ The BaByMERS also does not directly query about donor milk or milk sharing. Missing data for some questions were another limitation that undermined our ability to draw more definitive comparisons with the federal survey. Although this study did not identify many subgroups for which disagreement between the surveys was a disproportionate problem, the small group of Hispanic women was particularly likely to have high disagreement. Although respondents spoke English to participate, it is possible that some had difficulty with some questions. Validation of the BaByMERS in Spanish would be valuable.

Despite limitations, the study offered strengths in that we administered both surveys to the same respondents and in a random sequence. This enabled comparison of the two surveys and examination of ordering effects. The sample was also sociodemographically diverse and reflected the age distribution of children participating in several large national health surveys. Our method of interviewer administration also paralleled the methods of several national surveys. Finally, the study used the new BaByMERS, which had previously undergone initial validation using cognitive interviews.¹⁴ Although the BaByMERS is longer than the federal survey, it is shorter than other available tools that separately assess infant feeding modes.

Conclusion

This study offers further validation of a brief set of interviewer-administered questions, BaByMERS, to facilitate the collection of quality data recalled about infant feeding and lactation history in both federal and nonfederal studies. The BaByMERS appears to offer several advantages over questions commonly used in federal child health surveys. Although the BaByMERS is longer than the federal survey module, the BaByMERS generates additional estimates about infant feeding and lactation that are useful for research and public health surveillance. Further field testing with other populations and contexts will help confirm its utility for future research.

Footnotes

Acknowledgments

We thank The Ohio State University Center for Clinical and Translational Science; and Hanna Schlaack, Rachel Klenzman, Erin Shafer, and Thalia Cronin of the Center for Biobehavioral Health; and Kelly Crawford, Karissa Madison, and Christine Bally of the Division of Ambulatory Pediatrics of Nationwide Children's Hospital. The authors have no consultantships, honoraria, stock ownership, equity interests, arrangements regarding patents, or other vested interests related to the content of this article.

Funding Sources

The project described was supported by the Grant 1R03SH000048, funded by the Centers for Disease Control and Prevention and Award Number Grant UL1TR001070 from the National Center for Advancing Translational Sciences. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Centers for Disease Control and Prevention, the National Center for Advancing Translational Sciences, the National Institutes of Health, or the Department of Health and Human Services. The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review or approval of the article; and decision to submit the article for publication.

Disclosure Statement

No competing financial interests exist.

References

National Immunization Survey, National Center for Health Statistics, Centers for Disease Control and Prevention. 2017. Available at https://www.cdc.gov/vaccines/imz-managers/index.html (accessed April 1, 2019).

C.D.C. Breastfeeding Report Card 2018. Centers for Disease Control and Prevention. 2018. Available at https://wwwcdcgov/breastfeeding/pdf/2016breastfeedingreportcardpdf (accessed April 1, 2019).

National Survey of Children's Health 2016/17. Child and Adolescent Health Measurement Initiative, Data Resource Center on Child and Adolescent Health website. 2018. Available at www.childhealthdata.org (accessed December 6, 2018)

NHANES. National Center for Health Statistics. National Health and Nutrition Examination Survey. 2015–2016. Available at www.cdc.gov/nchs/nhanes.htm (accessed April 1, 2019).

Binns

, Win

, Zhao

, et al. Trends in the expression of breastmilk 1993–2003. Breastfeed Rev, 2006; 14:5–9.

Labiner-Wolfe

, Fein

, Shealy

, et al. Prevalence of breast milk expression and associated factors. Pediatrics, 2008; 122(Supplement 2):S63–S68.

Keim

, Boone

, Oza-Frank

, et al. Pumping milk without ever feeding at the breast in the Moms2Moms Study. Breastfeed Med, 2017; 12:422–429.

Rasmussen

, Felice

, O'Sullivan

, et al. The meaning of “Breastfeeding” is changing and so must our language about it. Breastfeed Med, 2017; 12:510–514.

US Department of Health and Human Services. 2015–2017 National Survey of Family Growth [Public use data tape documentation]. Hyattsville, MD, 2017.

10.

Soto-Ramirez

, Karmaus

, Zhang

, et al. Modes of infant feeding and the occurrence of coughing/wheezing in the first year of life. J Hum Lact, 2013; 29:71–80.

11.

, Magadia

, Fein

, et al. Risk of bottle-feeding for rapid weight gain during the first year of life. Arch Pediatr Adolesc Med, 2012; 166:431–436.

12.

Boone

, Geraghty

, Keim

. Feeding at the breast and expressed milk feeding: Associations with otitis media and diarrhea in infants. J Pediatr, 2016; 174:118–125.

13.

Klopp

, Vehling

, Becker

, et al. Modes of infant feeding and the risk of childhood asthma: A prospective birth cohort study. J Pediatr, 2017; 190:192–199.

14.

Keim

, Smith

, Boone

, et al. Cognitive testing of the Brief Breastfeeding and Milk Expression Recall Survey. Breastfeed Med, 2018; 13:60–66.

15.

Geraghty

, Rasmussen

. Redefining “breastfeeding” initiation and duration in the age of breastmilk pumping. Breastfeed Med, 2010; 5:135–137.

16.

Harris

, Taylor

, Thielke

, et al. Research electronic data capture (REDCap)—A metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform, 2009; 42:377–381.

17.

U.S. Department of Health and Human Services Office of Disease Prevention and Health Promotion. Healthy People 2020. 2014. Available at www.healthypeople.gov (accessed April 1, 2019).

18.

Victora

, Bahl

, Barros

AJD

, et al. Breastfeeding in the 21st century: Epidemiology, mechanisms, and lifelong effect. Lancet, 2016; 387:475–490.

19.

Bartok

. Babies fed breastmilk by breast versus by bottle: A pilot study evaluating early growth patterns. Breastfeed Med, 2011; 6:117–124.

20.

Azad

, Vehling

, Chan

, et al. Infant feeding and weight gain: Separating Breast milk from breastfeeding and formula from food. Pediatrics, 2018; 142:pii: e20181092.

21.

Fein

, Labiner-Wolfe

, Shealy

, et al. Infant feeding practices study II: Study methods. Pediatrics, 2008; 122:S28–S35.

22.

Noel-Weiss

, Taljaard

, Kujawa-Myles

. Breastfeeding and lactation research: Exploring a tool to measure infant feeding patterns. Int Breastfeed J, 2014; 9:5.

23.

O'Sullivan

, Rasmussen

. Development, construct validity, and reliability of the questionnaire on infant feeding: A tool for measuring contemporary infant-feeding behaviors. J Acad Nutr Diet, 2017; 117:1983–1990.e4.