- Email: [email protected]
Educational Intervention to Modify Bottle-feeding Behaviors among Formula-feeding Mothers in the WIC Program: Impact on Infant Formula Intake and Weight Gain
RESEARCH BRIEF
Educational Intervention to Modify Bottle-feeding Behaviors among Formula-feeding Mothers in the WIC Program: Impact on Infant Formula Intake and Weight Gain Katherine F. Kavanagh, PhD, RD1; Roberta J. Cohen, PhD2; M. Jane Heinig, PhD, IBCLC2; Kathryn G. Dewey, PhD2 ABSTRACT Objective: Formula-fed infants gain weight faster than breastfed infants. This study evaluated whether encouraging formula-feeding caregivers to be sensitive to infant satiety cues would alter feeding practices and reduce infant formula intake and weight gain. Design: Double-blind, randomized educational intervention, with intake and growth measured before (at 1 to 2 months) and after (4 to 5 months) the intervention. Setting: Women, Infants, and Children (WIC) clinics in Sacramento, California. Participants: 836 caregivers of young infants were screened; 214 were eligible, and 104 agreed to participate. Intervention: Intervention subjects received education promoting awareness of satiety cues and discouraging bottles containing more than 6 ounces before 4 months of age; intervention and control groups received education regarding introduction and feeding of solid food after 4 months of age. Main Outcome Measures: Formula intake (mL/24 hours) and weight gain (g/week). Analysis: Differences between groups evaluated using 2-way analysis of covariance (ANCOVA). Results: Sixty-one subjects completed baseline records, 44 attended class, and 38 completed the study. Despite a positive response to the educational intervention, there was no change in bottlefeeding behaviors (formula intake at 4 to 5 months was more than 1100 mL/day in both groups). Infant growth in the intervention group was greater than in the control group (P ⬍ .01), contrary to the hypothesis. Conclusions and Implications: The intervention improved knowledge of the key messages, but further research is needed to understand barriers to modifying bottle-feeding behaviors. Key Words: infant nutrition, formula-feeding, WIC, formula intake, education ( J Nutr Educ Behav. 2008;40:244-250)
INTRODUCTION One of the key factors associated with child overweight is a rapid rate of weight gain during infancy.1-3 Formula-fed infants consume more energy and gain weight more rapidly than breastfed infants, even during the first few months of 1
University of Tennessee, Knoxville, Tennessee University of California, Davis, Davis, California This study was supported by the Economic Research Program (ERS) Small Grants Program and the Food Stamp Nutrition Education Program (FSNEP) of California.
2
Address for correspondence: Katie F. Kavanagh, The University of Tennessee, 229 Jessie Harris Building, 1215 West Cumberland, Knoxville, TN 37996-1920; Tel: (865) 974-6250; Fax: (865) 974-3491E-mail: [email protected]
©2008 SOCIETY FOR NUTRITION EDUCATION doi: 10.1016/j.jneb.2007.01.002
life.4 Although formula-fed and breastfed infants are generally similar in percentage body fat during the first 6 months, formula-fed infants become significantly fatter than breastfed infants by the age of 12 months.5-6 Recent evidence indicates that there is a long-term effect of infant feeding on body fatness; children and adolescents who were breastfed are 20% to 30% less likely to be overweight than children who were formula fed.7 The mechanisms underlying these differences are not well understood. One possibility is that the composition of infant formulas has a stimulatory effect on intake and growth, although recent data from one of the investigators’ own studies suggest that neither the protein content or quality nor the potential renal solute load of formula is the
Journal of Nutrition Education and Behavior ● Volume 40, Number 4, July/August 2008
trigger.8 Another possibility is that the practice of bottlefeeding, not the composition of the milk in the bottle, is more important. It has been hypothesized that infants are born with the ability to self-regulate their energy intake.9 Breastfed infants may be better able to maintain this selfregulation over time because they are in control of when to terminate a feed. However, a caregiver who formula-feeds and ignores or is unaware of these cues may encourage the infant to take more formula even after the infant is satisfied. Caregivers may be motivated to empty the bottle in the mistaken belief that rapid weight gain is desirable,10,11 to soothe or encourage the infant to sleep longer,12,13 or to avoid wasting formula. Repeated overfeeding of formula-fed infants may cause them to lose the ability to precisely self-regulate energy intake, which would explain the observation that differences in intake between breastfed and formula-fed infants widen with age between 1 and 5 months, deviating significantly by 3 months of age.8 The objective of the current research was to determine whether an educational intervention to advise formulafeeding mothers to adopt “responsive” bottle-feeding practices (ie, avoid encouraging the infant to empty the bottle after the infant shows signs of satiety) and to limit the amount of formula initially offered at each feed to 6 ounces or less would result in (1) adoption of these practices, and (2) a lower volume of formula consumed at 4 months and less rapid weight gain from baseline to approximately 4 months.
DESCRIPTION OF INTERVENTION The study was a double-blind, randomized controlled trial of nutrition education for formula-feeding caregivers in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC). The control group received general guidance on infant feeding, and the intervention group received the same guidance plus specific advice to avoid making larger amounts of formula than necessary and to stop feeding when the infant first showed signs of satiety. Anthropometrists were masked to group assignment, and caregivers were not told that the main difference between the 2 groups was the advice on bottle-feeding methods. Caregivers whose infants were 3 to 10 weeks of age were recruited from 2 WIC clinics in Sacramento, California. Selection criteria included: (1) exclusively formula feeding, (2) birth weight at least 2500 grams and no chronic illness, (3) phone in the home or readily accessible, (4) caregiver planned to remain in the area until the infant was at least 4 months of age, (5) caregiver spoke English or Spanish, (6) caregiver willing to delay introduction of solid food until 4 months or later, and (7) infant not in foster care. The target sample size was 125 per group, calculated to allow detection of a difference in formula intake between groups of 110 mL (an effect size of 0.4), with a P-value of .05 and  ⫽ .20, and allowing for up to 25% attrition. After screening, eligible caregivers completed a baseline
245
2-day formula intake record. They recorded, to the nearest half ounce, how much was prepared and how much was left at each feed. A baseline interview was conducted to collect information on sociodemographic characteristics, breastfeeding history, infant feeding practices, and time the infant was in the care of others. To assess attitudes toward control of infant feeding, caregivers were asked to respond to several statements adapted from a questionnaire modified by Matthew Gillman (personal communication, 2005) from earlier work by Leann Birch (unpublished data, 2005), both recognized researchers in the area of parental control of child food intake. Research staff assessed the quality of the intake records by looking for large time gaps between feedings, implausible total volumes of formula (very small or very large) offered during the 48-hour period, or very little variability in the amount of formula left over after each feeding. Records that were suspected to be inaccurate were reviewed by the supervisor and, if found to be questionable, participants were asked to repeat either all (48 hours) or part (24 hours) of the record. Subjects were then stratified by infant sex and maternal language (English or Spanish) and randomized to attend the control or the intervention class. Intervention messages were generated and tested via focus groups in both English and Spanish with a similar WIC population in a neighboring county. The intervention and control educational modules were based on the Experiential Learning Cycle (ELC) of Kolb.14 Both classes covered general guidelines for infant feeding, including the appropriate age of introduction of solid food, safe preparation and feeding of solid food, responsive feeding practices when feeding solid food, and strategies to optimize nutrient adequacy. In the control group, additional information was provided on low-cost ways of providing nutritionally balanced meals to infants once they begin receiving solid food (after the completion of the study). Educators were trained to teach both control and intervention classes. Participants attended one 45- to 60-minute class, which directly replaced an existing WIC class. In the intervention group, the key messages were as follows:
1. Be aware of and responsive to the infant’s satiety cues. Satiety cues occurring early (slower sucking, getting sleepy, beginning to lose interest in the feed) and late (turning away, dribbling milk, biting the nipple, falling asleep, spitting up) were discussed, and caregivers were encouraged to stop feeding when the infant began demonstrating early cues. Educators distinguished among satisfaction, fullness, and uncomfortable fullness. This demonstration included a brief video puppet show in which the “infant” repeatedly exhibits satiety cues that should signal the “mother” to stop the feed. However, the “mother” ignores them and relies only on the amount of formula consumed to determine when the feed should end.
246
Kavanaugh et al/INTERVENTION TO MODIFY BOTTLE-FEEDING BEHAVIORS AMONG WIC MOTHERS
2. Prepare no more than 6 ounces at a feed. Breastfed infants rarely take more than 6 ounces at a feed. Therefore, caregivers were encouraged to prepare bottles containing no more than 6 ounces and to prepare additional formula if the infant was still hungry. Food models of a walnut, a medium-sized apple, and a large orange were used to demonstrate infant stomach capacity. The walnut corresponded to the infant’s stomach when empty, the apple to when the infant may reach satiety (approximately 5 to 6 ounces), and the orange to when an infant may be feeling overfull (about 8 ounces). In addition, the intervention group received a handout developed collaboratively with Dr. Harvey Karp explaining his techniques for calming a fussy baby,15 as well as a copy of his video, “The Happiest Baby on the Block.”
EVALUATION OF INTERVENTION Maternal height and weight and infant weight and length were measured after the class. Anthropometrists were trained and standardized in accordance with World Health Organization guidelines.16 All subjects who attended the class were followed for no less than 2 months after the class. Additional formula intake records were completed at 2 weeks post-class and around 3.5 months of age. To assess reactions to the educational messages, all participants completed an exit interview conducted by a research assistant who did not teach the classes. Infant anthropometry was repeated around 4 months of age. Caregivers received a $20 gift card after completing each of the first 2 intake records and a $30 gift card after completing the study. The study protocol was approved by the Institutional Review Board of the University of California, Davis. Statistical analysis was performed using SAS for Windows (version 8.02, SAS Institute, Cary, NC). Data analysis included examination of each variable for normality, descriptive statistics, and initial comparisons of intervention groups using Student t tests and chi-square tests. Differences in outcomes between groups were evaluated using 2-way analysis of covariance, controlling for baseline status (eg, formula intake or baseline anthropometric value), infant sex, age at entry into the study, and length of time in the study. Of the 836 caregivers screened, 214 were eligible, and 104 were willing to participate in the study. The main reasons for refusal were lack of time or interest (52.6%) and “access barriers” (21.7%), which included lack of transportation to the nutrition education class, uncertainty in ability to attend class, and family or personal problems. Of the 104 caregivers who agreed to participate, 101 completed the baseline questionnaire, and 61 completed the first formula intake record satisfactorily. The remainder (n ⫽ 43) did not complete the baseline intake record and therefore were not included in the randomized trial. In most of these cases, the research staff was never able to reconnect with the caregivers, even after multiple attempts. Of the 61
randomized, 17 did not attend the nutrition education class, even after repeated rescheduling. Thus, 44 caregivers attended the class, of whom 40 completed the final formula intake record and 38 attended the final measurement session. Of the 101 subjects completing the baseline questionnaire, there were no significant differences in characteristics between those who attended the class (n ⫽ 44) and those who did not (n ⫽ 57). Of the caregivers completing the final intake record (n ⫽ 40), there were no significant differences between intervention and control groups in maternal age (24.7 ⫾ 4.9 years intervention; 26.1 ⫾ 4.5 years control), education (10.7 ⫾ 3.2 years intervention; 10.8 ⫾ 3.4 years control), body mass index (28.2 ⫾ 4.7 kg/m2 intervention; 29.2 ⫾ 5.6 kg/m2 control), number of children (2.1 ⫾ 1.3 intervention; 2.8 ⫾ 1.2 control), infant birth weight (3429 ⫾ 433 grams intervention; 3578 ⫾ 568 grams control), or infant gender (52% of intervention group male; 47% of control group male). In addition, there were no significant differences between groups in previous breastfeeding experience or duration, ethnicity, working outside the home, or time the infant spent in the care of others during the study period. Results are shown in the Table. Formula intake did not differ significantly between intervention and control groups at baseline, nor at the midpoint or at the end of the study after controlling for age at baseline intake, baseline intake, infant sex, birth weight, and time in the study. There were no significant differences in infant weight or length at baseline after controlling for age and sex. At the end of the study, infants in the intervention group were heavier and longer than those in the control group, after controlling for age at measurement, sex, baseline anthropometric status, and time in study. Groups also did not differ significantly in bottle-feeding behaviors at baseline or at the final intake record. Bottle-emptying increased in both groups over time (from about 50% to about 60% of feedings), as did the use of bottles containing more than 6 ounces (from ⬍ 5% to about 17% of feedings). At baseline, about 10% to 14% of caregivers said that they often had to encourage the child to take more formula, and 16% to 24% felt that they had to be sure the infant finished the bottle. By contrast, about half felt that they had to be careful not to feed the child too much. There was little change in the first 2 responses by the end of the study in either group, but the percentage of caregivers who felt that they had to be careful not to feed the child too much increased from 53% to 84% in the intervention group, compared to a change from 53% to 57% in the control group (P ⫽ .07). All 40 caregivers found the class information to be helpful and agreed with the key messages. Fifty percent of the intervention group and 75% of the control group stated that the information in class was new to them. The intervention group (n ⫽ 21) reported that the most helpful information was (1) learning infant fullness cues (32%), (2)
Table. Comparison of Formula Intake, Weight, Length, and Bottle-feeding Outcomes of Intervention and Control Groups
Variables Formula Intake
a
Midpoint formula intake (n ⫽ 39) (mL/ 24 h) Final formula intake (mL/24 h) Change in intake from baseline to end of study (mL/24 h) Weight and Lengthf Baseline weight, g Final weight, g Gain per week, g Baseline length, cm Final length, cm Gain in length per week, cm Bottle-feeding variablesj Percentage of bottles emptied at baseline Percentage of bottles emptied at final intake record Change in percentage of bottles emptied
Adjusted Mean (SE)
Intervention (n ⴝ 19) 786 (181) 598-1281 911 (296) 628-1985 1119 (297) 622-1996
800 (43.8) 918 (43.3) 1128 (67.2)
Unadjusted Mean (SD) range
Adjusted Mean (SE)
P-value (from ANCOVA)
785 (41.6)
NSb
908 (42.2)
NSc
1107 (63.8)
NSd
315 (63.8)
NSe
4987 (111)
NSg
6758 (106)
.006h
156.1 (9.5) 56.3 (0.4)
.008h NSg
63.3 (0.2)
.02i
0.63 (0.02)
.045i
50.8 (6.1)
NSk
60.6 (5.4)
NSl
⫹10.3 (5.4)
NSl
Control (n ⴝ 21) 799 (195) 504-1317 915 (330) 537-1895 1115 (466) 674-2565
335 (67.2) Intervention (n ⴝ 18) 5157 (821) 3970-6570 7262 (683) 6091-8431 56.6 (2.8) 52.2-61.0 64.2 (1.4) 60.8-66.4
4998 (118) 7214 (112) 195.3 (10.0) 56.1 (0.4) 64.2 (0.3)
0.70 (0.02) Intervention (n ⴝ 19) 50.4 (27.2) 5.3-100 60.7 (28.6) 0-100
49.9 (6.5) 60.6 (5.6) ⫹10.2 (5.6)
Control (n ⴝ 20) 4844 (818) 3676-6670 6715 (718) 5819-8934 55.8 (2.6) 50.8-60.0 63.3 (1.9) 59.8-67.0
Control (n ⴝ 21) 50.3 (28.4) 0-100 60.5 (24.3) 21.4-96.2
Journal of Nutrition Education and Behavior ● Volume 40, Number 4, July/August 2008
Baseline formula intake (mL/24 h)
Unadjusted Mean (SD) range
247
248
Number (%) emptying bottle at ⬎ 50% of feedings at baseline Number (%) emptying bottle at ⬎ 50% of feedings at final intake record Percentage of bottles ⬎ 6 oz offered at baseline Percentage of bottles ⬎ 6 oz offered at final intake record Change in percentage of bottles ⬎ 6 oz offered
8 (42%)
11 (52%)
NS
12 (63%)
13 (62%)
NS
2.5 (9.7) 0-42.1 20.0 (32.4) 0-100
1.7 (2.8) 23.6 (6.3) ⫹20.3 (6.3)
4.1 (14.0) 0-60 15.2 (30.3) 0-100
4.8 (2.6)
NSk
11.9 (6.0)
NSm
⫹8.6 (6.0)
NSm
a. For each infant, 24 h formula intake was calculated by dividing the total oz consumed during the 48 h period by the time interval (in hours) between the beginning of the first feeding and the beginning of the feeding that occurred just after the 48 h period, and then multiplying by 24. Of the 38 caregivers completing the study, 7 (18%) introduced some form of food or fluid other than formula before the infant was 4 mo of age, mostly teas or juice. b. Adjusted for age at baseline intake record, sex, and birth weight. c. Adjusted for age at midpoint intake record, baseline intake, sex, and birth weight. d. Adjusted for age at final intake record, baseline intake, sex, birth weight, and time in study. e. Adjusted for age at baseline, baseline intake, sex, birth weight, and time in study. f. Calculated as the change in weight or length per week of enrollment since initial measurement. g. Adjusted for age at baseline measurement and infant sex. h. Adjusted for age at baseline measurement, baseline weight, time in study and sex. i. Adjusted for age at baseline measurement, baseline length, time in study and sex. j. From each intake record, the proportion of bottles emptied (defined as ⬍1/2 oz remaining) was calculated for each day and the mean of the two values for each 48 h record was used. k. Adjusted for age at baseline intake record and sex. l. Adjusted for age at baseline intake record, time in study, baseline percentage of bottles emptied, and sex. m. Adjusted for age at baseline intake record, time in study, baseline percentage of bottles ⬎ 6 oz, and sex. SD indicates standard deviation; SE, standard error; ANCOVA, analysis of covariance; NS, not significant
Kavanaugh et al/INTERVENTION TO MODIFY BOTTLE-FEEDING BEHAVIORS AMONG WIC MOTHERS
Table. Comparison of Formula Intake, Weight, Length, and Bottle-feeding Outcomes of Intervention and Control Groups
Journal of Nutrition Education and Behavior ● Volume 40, Number 4, July/August 2008
strategies to limit bottles to no more than 6 ounces (16%), and (3) alternatives to offering the bottle (14%). Most of the caregivers (84%) found it easy to follow the advice of preparing no more than 6 ounces of formula at a time, often because the infant was not yet taking 6 ounces. Nearly all (95%) said that it was easy to follow the advice of watching for and acting on early cues of fullness, and 89% reported that informing friends and family about the educational messages was easy. Two-thirds indicated that friends, family, and day care were very supportive of what the caregivers were trying to do.
DISCUSSION This was the first randomized intervention trial designed to modify bottle-feeding behaviors that may contribute to excess weight gain during infancy and predispose to child overweight. Despite extensive efforts to develop and pretest appropriate educational messages and audiovisual aids using ELC techniques, the attempt was unsuccessful at reducing the frequency of bottle-emptying or the preparation of bottles with more than 6 ounces of formula. In both groups, bottle-emptying was relatively common both before (about 50% of feedings) and after (about 60% of feedings) the intervention. On the other hand, the use of bottles containing more than 6 ounces was relatively uncommon (⬍ 5% at baseline; about 17% at the end of the study), so there was less opportunity for influencing that behavior. Given that the intervention was apparently ineffective at modifying bottle-feeding behaviors, it is not surprising that there was no impact on formula intake. The significantly greater weight and length gain of infants in the intervention group, contrary to the hypothesis, was surprising. It is possible that these results reflect random error. It is unlikely that food other than formula contributed to this difference in growth, as there was no difference between groups in the type or frequency of other food reportedly offered (data not shown). Similarly, the differences cannot be explained by socioeconomic variables such as caregiver work or school status or time infant was in the care of others, as these variables did not differ by group. One possibility is that infants in the intervention group cried less often, were swaddled more, or were attended to more than those in the control group (perhaps owing to caregivers’ responses to the information on calming a fussy infant) and therefore expended less energy during the intervention period, though it is unknown whether any such difference in energy expenditure would be large enough to significantly affect growth. There are several potential explanations for the lack of behavioral change in the intervention group. First, it is possible that caregivers did not understand or remember the key messages sufficiently to implement the advice given, and these types of messages may require more support and repetition. In addition, though the ELC has been shown to be effective in similar populations,14 these activities were
249
originally designed for use in a group setting. Unfortunately, 95% of classes had to be conducted as one-on-one sessions because of no-shows. Therefore, the facilitation that occurs via group discussion and self-discovery of the usefulness of the messages, integral parts of the ELC learning process, may have been lost owing to the one-on-one nature of the classes. In addition, the implementation of the ELC was not assessed, other than observing classes and giving feedback. Despite this situation, response to the nutrition education classes and the key messages was positive, and caregivers overwhelmingly reported enjoying the intimate character of the classes. Caregivers appeared to value the information given in class and appeared willing to verbalize this knowledge to important people in their lives. In addition, the intervention group’s attitudes about overfeeding appeared to change in a positive direction; the percentage of caregivers who felt that they had to be careful not to feed the child too much increased from 53% to 84% in the intervention group, whereas there was very little change in the control group. Another possibility is that the educational sessions, although convincing, did not adequately provide the specific tools or skills needed to achieve behavioral change. Almost half of caregivers in the intervention group said that the most interesting class topic was the information on ways to calm a fussy infant. Because feeding is often used to calm a fussy infant,12,13 and because caregivers were being instructed not to use food as the first strategy to comfort their infants, the intervention group was provided with a popular video of calming techniques. Originally, the plan was to show a somewhat shortened version of the video “The Happiest Baby on the Block” during each intervention class. However, owing to time constraints, educators instead sent the video home with caregivers, with the instruction to view it at home and to call with any questions. It is possible that, had there been sufficient time to view the video with each class and to respond to questions and reinforce techniques, educators could have addressed this concept more fully and thereby facilitated greater change in feeding practices. Educators did not ask if the videotape was viewed. During the class, early and late satiety cues were demonstrated with a video of puppets, but it might have been more effective to use video of real infants actually exhibiting these cues. Training caregivers to recognize the cues of their own infants may be an even more effective way to transmit both knowledge and skills for responsive feeding. Additionally, the curriculum may be more effective if offered as several shorter class sessions to build the caregivers’ skills and confidence. The timing of the intervention also may not have been ideal. It may be more appropriate to begin education on responsive feeding prenatally, as there are some indications that the unlearning of appetite regulation can begin shortly after birth.17 Extending the intervention past the point when the majority of caregivers would normally have begun to transition to bottle feeds of at least 6 ounces may also have been beneficial.
250
Kavanaugh et al/INTERVENTION TO MODIFY BOTTLE-FEEDING BEHAVIORS AMONG WIC MOTHERS
It is important to note that the sample size achieved was far lower than originally planned, as only about 18% of eligible subjects enrolled and successfully completed the study. The caregivers who attended class may have been more motivated, or had fewer time constraints, than the remainder of the eligible subjects, which may limit the generalizability of the findings.
IMPLICATIONS FOR RESEARCH AND PRACTICE The large increase in formula intake during the study, from about 800 mL per day to more than 1100 mL per day, is much greater than what is observed among breastfed infants during the same age interval8 and suggests that these infants were being overfed by the end of the study. Many caregivers felt that they were already using responsive feeding techniques, so although they agreed with the messages, they may not have felt the need to change their behavior. A certain percentage of “bottle-emptying” is to be expected, in other words, the infant will finish the bottle some of the time, without any prompting. What is unknown is whether infant-driven bottle-emptying can account for the relatively common occurrence of empty bottles observed (50% to 60%). It is likely that caregiver-driven bottleemptying accounts for a considerable proportion of this phenomenon, though it may not be recognized as a factor by the caregivers themselves. If this scenario is correct, what are the barriers to reducing bottle-emptying? In the focus group discussions conducted during the message development phase, the investigators found that caregivers were concerned with preventing their infants from crying too much and encouraging them to sleep through the night. The messages, which were focused on preventing child overweight (a longerterm outcome), may have been less compelling than the more short-term concerns of coping with a fussy infant. Caregivers use feeding as a strategy for soothing their infants,11-13,18,19 thus it may be difficult to change a behavior that they perceive as having an immediate benefit to them. In summary, the results of this study indicate that formula intakes of infants in this population are quite high— probably reflective of overfeeding—and that modifying bottle-feeding behaviors to prevent overfeeding is a challenging task. Further research is needed to understand the attitudes and life circumstances that are constraints to changing these behaviors.
ACKNOWLEDGMENT This study was supported by the Economic Research Program (ERS) Small Grants Program and the Food Stamp
Nutrition Education Program (FSNEP) of California. The authors wish to thank Dr. Harvey Karp for his assistance with this project.
REFERENCES 1. Ong KKL, Ahmed ML, Emmett PM, Preece MA, Dunger DB. Association between postnatal catch-up growth and obesity in childhood: prospective cohort study. BMJ. 2000;320:967-971. 2. Stettler N, Zemel BS, Kumanyika S, Stallings VA. Infant weight gain and childhood overweight status in a multicenter cohort study. Pediatrics. 2002;109:194-199. 3. Cameron N, Pettifor J, De Wet T, Norris S. The relationship of rapid weight gain in infancy to obesity and skeletal maturity in childhood. Obes Res. 2003;11:457-460. 4. Dewey KG. Growth characteristics of breast-fed compared to formulafed infants. Biol Neonate. 1998;74:94-105. 5. Dewey KG, Heinig MJ, Nommsen LA, Peerson JM, Lonnerdal B. Breastfed infants are leaner than formula-fed infants at 1 year of age: the DARLING study. Am J Clin Nutr. 1993;57:140-145. 6. Dewey KG, Peerson JM, Brown KH, et al. Growth of breast-fed infants deviates from current reference data: a pooled analysis of US, Canadian, and European data sets. Pediatrics. 1995;96:495-503. 7. Dewey KG. Is breastfeeding protective against child obesity? J Hum Lact. 2003;19:9-18. 8. Dewey KG, Nommsen-Rivers L, Lonnerdal B. Plasma insulin and insulin-releasing amino acid (IRAA) concentrations are higher in formula-fed than in breastfed infants at 5 mo of age [meeting abstract]. FASEB J. 2004;18(suppl):A843. 9. Dewey KG, Lonnerdal B. Infant self-regulation of breast milk intake. Acta Paediatr Scand. 1986;75:893-898. 10. Baughcum AE, Burklow KA, Deeks CM, Powers SW, Whitaker RC. Maternal feeding practices and childhood obesity. Arch Pediatr Adolesc Med. 1998;152:1010-1014. 11. Boyington JA, Johnson AA. Maternal perception of body size as a determinant of infant adiposity in an African-American community. J Natl Med Assoc. 2004;96:351-362. 12. Savage SA, Reilly JJ, Edwards CA, Durmin JV. Weaning practices in the Glasgow Longitudinal Infant Growth Study. Arch Dis Child. 1998;79:153-156. 13. Bentley M, Gavin L, Black MM, Teti L. Infant feeding practices of low-income, African-American, adolescent mothers: an ecological, multigenerational perspective. Soc Sci Med. 1999;49:1085-1100. 14. Baker CA, Jensen PJ, Kolb DA. Conversational Learning: An Experiential Approach to Knowledge Creation. Westport, Conn: Quorum Books; 2002. 15. Karp H. The Happiest Baby on the Block: The New Way to Calm Crying and Help Your Newborn Baby Sleep Longer. New York, NY: Bantam Dell; 2002. 16. de Onis M, Onyango AW, Van den Broeck J, Chumlea WC, Martorell R for the WHO Multicentre Growth Reference Study Group. Measurement and standardization protocols for anthropometry used in the construction of a new and international growth reference. Food Nutr Bull. 2004;25:S27-S36. 17. Dollberg S, Lahav S, Mimouni FB. A comparison of intakes of breast-fed and bottle-fed infants during the first two days of life. J Am Coll Nutr. 2001;20:209-211. 18. Hannon PR, Willis SK, Bishop-Townsend V, et al. African-American and Latina adolescent mothers infant feeding decisions and breastfeeding practices: a qualitative study. J Adol Health. 2000;26:399-407. 19. Bronner YL, Gross SM, Caufield L, et al. Early introduction of solid foods among urban African-American participants in WIC. J Am Diet Assoc. 1999;99:457-461.
Educational Intervention to Modify Bottle-feeding Behaviors among Formula-feeding Mothers in the WIC Program: Impact on Infant Formula Intake and Weight Gain Katherine F. Kavanagh, PhD, RD1; Roberta J. Cohen, PhD2; M. Jane Heinig, PhD, IBCLC2; Kathryn G. Dewey, PhD2 ABSTRACT Objective: Formula-fed infants gain weight faster than breastfed infants. This study evaluated whether encouraging formula-feeding caregivers to be sensitive to infant satiety cues would alter feeding practices and reduce infant formula intake and weight gain. Design: Double-blind, randomized educational intervention, with intake and growth measured before (at 1 to 2 months) and after (4 to 5 months) the intervention. Setting: Women, Infants, and Children (WIC) clinics in Sacramento, California. Participants: 836 caregivers of young infants were screened; 214 were eligible, and 104 agreed to participate. Intervention: Intervention subjects received education promoting awareness of satiety cues and discouraging bottles containing more than 6 ounces before 4 months of age; intervention and control groups received education regarding introduction and feeding of solid food after 4 months of age. Main Outcome Measures: Formula intake (mL/24 hours) and weight gain (g/week). Analysis: Differences between groups evaluated using 2-way analysis of covariance (ANCOVA). Results: Sixty-one subjects completed baseline records, 44 attended class, and 38 completed the study. Despite a positive response to the educational intervention, there was no change in bottlefeeding behaviors (formula intake at 4 to 5 months was more than 1100 mL/day in both groups). Infant growth in the intervention group was greater than in the control group (P ⬍ .01), contrary to the hypothesis. Conclusions and Implications: The intervention improved knowledge of the key messages, but further research is needed to understand barriers to modifying bottle-feeding behaviors. Key Words: infant nutrition, formula-feeding, WIC, formula intake, education ( J Nutr Educ Behav. 2008;40:244-250)
INTRODUCTION One of the key factors associated with child overweight is a rapid rate of weight gain during infancy.1-3 Formula-fed infants consume more energy and gain weight more rapidly than breastfed infants, even during the first few months of 1
University of Tennessee, Knoxville, Tennessee University of California, Davis, Davis, California This study was supported by the Economic Research Program (ERS) Small Grants Program and the Food Stamp Nutrition Education Program (FSNEP) of California.
2
Address for correspondence: Katie F. Kavanagh, The University of Tennessee, 229 Jessie Harris Building, 1215 West Cumberland, Knoxville, TN 37996-1920; Tel: (865) 974-6250; Fax: (865) 974-3491E-mail: [email protected]
©2008 SOCIETY FOR NUTRITION EDUCATION doi: 10.1016/j.jneb.2007.01.002
life.4 Although formula-fed and breastfed infants are generally similar in percentage body fat during the first 6 months, formula-fed infants become significantly fatter than breastfed infants by the age of 12 months.5-6 Recent evidence indicates that there is a long-term effect of infant feeding on body fatness; children and adolescents who were breastfed are 20% to 30% less likely to be overweight than children who were formula fed.7 The mechanisms underlying these differences are not well understood. One possibility is that the composition of infant formulas has a stimulatory effect on intake and growth, although recent data from one of the investigators’ own studies suggest that neither the protein content or quality nor the potential renal solute load of formula is the
Journal of Nutrition Education and Behavior ● Volume 40, Number 4, July/August 2008
trigger.8 Another possibility is that the practice of bottlefeeding, not the composition of the milk in the bottle, is more important. It has been hypothesized that infants are born with the ability to self-regulate their energy intake.9 Breastfed infants may be better able to maintain this selfregulation over time because they are in control of when to terminate a feed. However, a caregiver who formula-feeds and ignores or is unaware of these cues may encourage the infant to take more formula even after the infant is satisfied. Caregivers may be motivated to empty the bottle in the mistaken belief that rapid weight gain is desirable,10,11 to soothe or encourage the infant to sleep longer,12,13 or to avoid wasting formula. Repeated overfeeding of formula-fed infants may cause them to lose the ability to precisely self-regulate energy intake, which would explain the observation that differences in intake between breastfed and formula-fed infants widen with age between 1 and 5 months, deviating significantly by 3 months of age.8 The objective of the current research was to determine whether an educational intervention to advise formulafeeding mothers to adopt “responsive” bottle-feeding practices (ie, avoid encouraging the infant to empty the bottle after the infant shows signs of satiety) and to limit the amount of formula initially offered at each feed to 6 ounces or less would result in (1) adoption of these practices, and (2) a lower volume of formula consumed at 4 months and less rapid weight gain from baseline to approximately 4 months.
DESCRIPTION OF INTERVENTION The study was a double-blind, randomized controlled trial of nutrition education for formula-feeding caregivers in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC). The control group received general guidance on infant feeding, and the intervention group received the same guidance plus specific advice to avoid making larger amounts of formula than necessary and to stop feeding when the infant first showed signs of satiety. Anthropometrists were masked to group assignment, and caregivers were not told that the main difference between the 2 groups was the advice on bottle-feeding methods. Caregivers whose infants were 3 to 10 weeks of age were recruited from 2 WIC clinics in Sacramento, California. Selection criteria included: (1) exclusively formula feeding, (2) birth weight at least 2500 grams and no chronic illness, (3) phone in the home or readily accessible, (4) caregiver planned to remain in the area until the infant was at least 4 months of age, (5) caregiver spoke English or Spanish, (6) caregiver willing to delay introduction of solid food until 4 months or later, and (7) infant not in foster care. The target sample size was 125 per group, calculated to allow detection of a difference in formula intake between groups of 110 mL (an effect size of 0.4), with a P-value of .05 and  ⫽ .20, and allowing for up to 25% attrition. After screening, eligible caregivers completed a baseline
245
2-day formula intake record. They recorded, to the nearest half ounce, how much was prepared and how much was left at each feed. A baseline interview was conducted to collect information on sociodemographic characteristics, breastfeeding history, infant feeding practices, and time the infant was in the care of others. To assess attitudes toward control of infant feeding, caregivers were asked to respond to several statements adapted from a questionnaire modified by Matthew Gillman (personal communication, 2005) from earlier work by Leann Birch (unpublished data, 2005), both recognized researchers in the area of parental control of child food intake. Research staff assessed the quality of the intake records by looking for large time gaps between feedings, implausible total volumes of formula (very small or very large) offered during the 48-hour period, or very little variability in the amount of formula left over after each feeding. Records that were suspected to be inaccurate were reviewed by the supervisor and, if found to be questionable, participants were asked to repeat either all (48 hours) or part (24 hours) of the record. Subjects were then stratified by infant sex and maternal language (English or Spanish) and randomized to attend the control or the intervention class. Intervention messages were generated and tested via focus groups in both English and Spanish with a similar WIC population in a neighboring county. The intervention and control educational modules were based on the Experiential Learning Cycle (ELC) of Kolb.14 Both classes covered general guidelines for infant feeding, including the appropriate age of introduction of solid food, safe preparation and feeding of solid food, responsive feeding practices when feeding solid food, and strategies to optimize nutrient adequacy. In the control group, additional information was provided on low-cost ways of providing nutritionally balanced meals to infants once they begin receiving solid food (after the completion of the study). Educators were trained to teach both control and intervention classes. Participants attended one 45- to 60-minute class, which directly replaced an existing WIC class. In the intervention group, the key messages were as follows:
1. Be aware of and responsive to the infant’s satiety cues. Satiety cues occurring early (slower sucking, getting sleepy, beginning to lose interest in the feed) and late (turning away, dribbling milk, biting the nipple, falling asleep, spitting up) were discussed, and caregivers were encouraged to stop feeding when the infant began demonstrating early cues. Educators distinguished among satisfaction, fullness, and uncomfortable fullness. This demonstration included a brief video puppet show in which the “infant” repeatedly exhibits satiety cues that should signal the “mother” to stop the feed. However, the “mother” ignores them and relies only on the amount of formula consumed to determine when the feed should end.
246
Kavanaugh et al/INTERVENTION TO MODIFY BOTTLE-FEEDING BEHAVIORS AMONG WIC MOTHERS
2. Prepare no more than 6 ounces at a feed. Breastfed infants rarely take more than 6 ounces at a feed. Therefore, caregivers were encouraged to prepare bottles containing no more than 6 ounces and to prepare additional formula if the infant was still hungry. Food models of a walnut, a medium-sized apple, and a large orange were used to demonstrate infant stomach capacity. The walnut corresponded to the infant’s stomach when empty, the apple to when the infant may reach satiety (approximately 5 to 6 ounces), and the orange to when an infant may be feeling overfull (about 8 ounces). In addition, the intervention group received a handout developed collaboratively with Dr. Harvey Karp explaining his techniques for calming a fussy baby,15 as well as a copy of his video, “The Happiest Baby on the Block.”
EVALUATION OF INTERVENTION Maternal height and weight and infant weight and length were measured after the class. Anthropometrists were trained and standardized in accordance with World Health Organization guidelines.16 All subjects who attended the class were followed for no less than 2 months after the class. Additional formula intake records were completed at 2 weeks post-class and around 3.5 months of age. To assess reactions to the educational messages, all participants completed an exit interview conducted by a research assistant who did not teach the classes. Infant anthropometry was repeated around 4 months of age. Caregivers received a $20 gift card after completing each of the first 2 intake records and a $30 gift card after completing the study. The study protocol was approved by the Institutional Review Board of the University of California, Davis. Statistical analysis was performed using SAS for Windows (version 8.02, SAS Institute, Cary, NC). Data analysis included examination of each variable for normality, descriptive statistics, and initial comparisons of intervention groups using Student t tests and chi-square tests. Differences in outcomes between groups were evaluated using 2-way analysis of covariance, controlling for baseline status (eg, formula intake or baseline anthropometric value), infant sex, age at entry into the study, and length of time in the study. Of the 836 caregivers screened, 214 were eligible, and 104 were willing to participate in the study. The main reasons for refusal were lack of time or interest (52.6%) and “access barriers” (21.7%), which included lack of transportation to the nutrition education class, uncertainty in ability to attend class, and family or personal problems. Of the 104 caregivers who agreed to participate, 101 completed the baseline questionnaire, and 61 completed the first formula intake record satisfactorily. The remainder (n ⫽ 43) did not complete the baseline intake record and therefore were not included in the randomized trial. In most of these cases, the research staff was never able to reconnect with the caregivers, even after multiple attempts. Of the 61
randomized, 17 did not attend the nutrition education class, even after repeated rescheduling. Thus, 44 caregivers attended the class, of whom 40 completed the final formula intake record and 38 attended the final measurement session. Of the 101 subjects completing the baseline questionnaire, there were no significant differences in characteristics between those who attended the class (n ⫽ 44) and those who did not (n ⫽ 57). Of the caregivers completing the final intake record (n ⫽ 40), there were no significant differences between intervention and control groups in maternal age (24.7 ⫾ 4.9 years intervention; 26.1 ⫾ 4.5 years control), education (10.7 ⫾ 3.2 years intervention; 10.8 ⫾ 3.4 years control), body mass index (28.2 ⫾ 4.7 kg/m2 intervention; 29.2 ⫾ 5.6 kg/m2 control), number of children (2.1 ⫾ 1.3 intervention; 2.8 ⫾ 1.2 control), infant birth weight (3429 ⫾ 433 grams intervention; 3578 ⫾ 568 grams control), or infant gender (52% of intervention group male; 47% of control group male). In addition, there were no significant differences between groups in previous breastfeeding experience or duration, ethnicity, working outside the home, or time the infant spent in the care of others during the study period. Results are shown in the Table. Formula intake did not differ significantly between intervention and control groups at baseline, nor at the midpoint or at the end of the study after controlling for age at baseline intake, baseline intake, infant sex, birth weight, and time in the study. There were no significant differences in infant weight or length at baseline after controlling for age and sex. At the end of the study, infants in the intervention group were heavier and longer than those in the control group, after controlling for age at measurement, sex, baseline anthropometric status, and time in study. Groups also did not differ significantly in bottle-feeding behaviors at baseline or at the final intake record. Bottle-emptying increased in both groups over time (from about 50% to about 60% of feedings), as did the use of bottles containing more than 6 ounces (from ⬍ 5% to about 17% of feedings). At baseline, about 10% to 14% of caregivers said that they often had to encourage the child to take more formula, and 16% to 24% felt that they had to be sure the infant finished the bottle. By contrast, about half felt that they had to be careful not to feed the child too much. There was little change in the first 2 responses by the end of the study in either group, but the percentage of caregivers who felt that they had to be careful not to feed the child too much increased from 53% to 84% in the intervention group, compared to a change from 53% to 57% in the control group (P ⫽ .07). All 40 caregivers found the class information to be helpful and agreed with the key messages. Fifty percent of the intervention group and 75% of the control group stated that the information in class was new to them. The intervention group (n ⫽ 21) reported that the most helpful information was (1) learning infant fullness cues (32%), (2)
Table. Comparison of Formula Intake, Weight, Length, and Bottle-feeding Outcomes of Intervention and Control Groups
Variables Formula Intake
a
Midpoint formula intake (n ⫽ 39) (mL/ 24 h) Final formula intake (mL/24 h) Change in intake from baseline to end of study (mL/24 h) Weight and Lengthf Baseline weight, g Final weight, g Gain per week, g Baseline length, cm Final length, cm Gain in length per week, cm Bottle-feeding variablesj Percentage of bottles emptied at baseline Percentage of bottles emptied at final intake record Change in percentage of bottles emptied
Adjusted Mean (SE)
Intervention (n ⴝ 19) 786 (181) 598-1281 911 (296) 628-1985 1119 (297) 622-1996
800 (43.8) 918 (43.3) 1128 (67.2)
Unadjusted Mean (SD) range
Adjusted Mean (SE)
P-value (from ANCOVA)
785 (41.6)
NSb
908 (42.2)
NSc
1107 (63.8)
NSd
315 (63.8)
NSe
4987 (111)
NSg
6758 (106)
.006h
156.1 (9.5) 56.3 (0.4)
.008h NSg
63.3 (0.2)
.02i
0.63 (0.02)
.045i
50.8 (6.1)
NSk
60.6 (5.4)
NSl
⫹10.3 (5.4)
NSl
Control (n ⴝ 21) 799 (195) 504-1317 915 (330) 537-1895 1115 (466) 674-2565
335 (67.2) Intervention (n ⴝ 18) 5157 (821) 3970-6570 7262 (683) 6091-8431 56.6 (2.8) 52.2-61.0 64.2 (1.4) 60.8-66.4
4998 (118) 7214 (112) 195.3 (10.0) 56.1 (0.4) 64.2 (0.3)
0.70 (0.02) Intervention (n ⴝ 19) 50.4 (27.2) 5.3-100 60.7 (28.6) 0-100
49.9 (6.5) 60.6 (5.6) ⫹10.2 (5.6)
Control (n ⴝ 20) 4844 (818) 3676-6670 6715 (718) 5819-8934 55.8 (2.6) 50.8-60.0 63.3 (1.9) 59.8-67.0
Control (n ⴝ 21) 50.3 (28.4) 0-100 60.5 (24.3) 21.4-96.2
Journal of Nutrition Education and Behavior ● Volume 40, Number 4, July/August 2008
Baseline formula intake (mL/24 h)
Unadjusted Mean (SD) range
247
248
Number (%) emptying bottle at ⬎ 50% of feedings at baseline Number (%) emptying bottle at ⬎ 50% of feedings at final intake record Percentage of bottles ⬎ 6 oz offered at baseline Percentage of bottles ⬎ 6 oz offered at final intake record Change in percentage of bottles ⬎ 6 oz offered
8 (42%)
11 (52%)
NS
12 (63%)
13 (62%)
NS
2.5 (9.7) 0-42.1 20.0 (32.4) 0-100
1.7 (2.8) 23.6 (6.3) ⫹20.3 (6.3)
4.1 (14.0) 0-60 15.2 (30.3) 0-100
4.8 (2.6)
NSk
11.9 (6.0)
NSm
⫹8.6 (6.0)
NSm
a. For each infant, 24 h formula intake was calculated by dividing the total oz consumed during the 48 h period by the time interval (in hours) between the beginning of the first feeding and the beginning of the feeding that occurred just after the 48 h period, and then multiplying by 24. Of the 38 caregivers completing the study, 7 (18%) introduced some form of food or fluid other than formula before the infant was 4 mo of age, mostly teas or juice. b. Adjusted for age at baseline intake record, sex, and birth weight. c. Adjusted for age at midpoint intake record, baseline intake, sex, and birth weight. d. Adjusted for age at final intake record, baseline intake, sex, birth weight, and time in study. e. Adjusted for age at baseline, baseline intake, sex, birth weight, and time in study. f. Calculated as the change in weight or length per week of enrollment since initial measurement. g. Adjusted for age at baseline measurement and infant sex. h. Adjusted for age at baseline measurement, baseline weight, time in study and sex. i. Adjusted for age at baseline measurement, baseline length, time in study and sex. j. From each intake record, the proportion of bottles emptied (defined as ⬍1/2 oz remaining) was calculated for each day and the mean of the two values for each 48 h record was used. k. Adjusted for age at baseline intake record and sex. l. Adjusted for age at baseline intake record, time in study, baseline percentage of bottles emptied, and sex. m. Adjusted for age at baseline intake record, time in study, baseline percentage of bottles ⬎ 6 oz, and sex. SD indicates standard deviation; SE, standard error; ANCOVA, analysis of covariance; NS, not significant
Kavanaugh et al/INTERVENTION TO MODIFY BOTTLE-FEEDING BEHAVIORS AMONG WIC MOTHERS
Table. Comparison of Formula Intake, Weight, Length, and Bottle-feeding Outcomes of Intervention and Control Groups
Journal of Nutrition Education and Behavior ● Volume 40, Number 4, July/August 2008
strategies to limit bottles to no more than 6 ounces (16%), and (3) alternatives to offering the bottle (14%). Most of the caregivers (84%) found it easy to follow the advice of preparing no more than 6 ounces of formula at a time, often because the infant was not yet taking 6 ounces. Nearly all (95%) said that it was easy to follow the advice of watching for and acting on early cues of fullness, and 89% reported that informing friends and family about the educational messages was easy. Two-thirds indicated that friends, family, and day care were very supportive of what the caregivers were trying to do.
DISCUSSION This was the first randomized intervention trial designed to modify bottle-feeding behaviors that may contribute to excess weight gain during infancy and predispose to child overweight. Despite extensive efforts to develop and pretest appropriate educational messages and audiovisual aids using ELC techniques, the attempt was unsuccessful at reducing the frequency of bottle-emptying or the preparation of bottles with more than 6 ounces of formula. In both groups, bottle-emptying was relatively common both before (about 50% of feedings) and after (about 60% of feedings) the intervention. On the other hand, the use of bottles containing more than 6 ounces was relatively uncommon (⬍ 5% at baseline; about 17% at the end of the study), so there was less opportunity for influencing that behavior. Given that the intervention was apparently ineffective at modifying bottle-feeding behaviors, it is not surprising that there was no impact on formula intake. The significantly greater weight and length gain of infants in the intervention group, contrary to the hypothesis, was surprising. It is possible that these results reflect random error. It is unlikely that food other than formula contributed to this difference in growth, as there was no difference between groups in the type or frequency of other food reportedly offered (data not shown). Similarly, the differences cannot be explained by socioeconomic variables such as caregiver work or school status or time infant was in the care of others, as these variables did not differ by group. One possibility is that infants in the intervention group cried less often, were swaddled more, or were attended to more than those in the control group (perhaps owing to caregivers’ responses to the information on calming a fussy infant) and therefore expended less energy during the intervention period, though it is unknown whether any such difference in energy expenditure would be large enough to significantly affect growth. There are several potential explanations for the lack of behavioral change in the intervention group. First, it is possible that caregivers did not understand or remember the key messages sufficiently to implement the advice given, and these types of messages may require more support and repetition. In addition, though the ELC has been shown to be effective in similar populations,14 these activities were
249
originally designed for use in a group setting. Unfortunately, 95% of classes had to be conducted as one-on-one sessions because of no-shows. Therefore, the facilitation that occurs via group discussion and self-discovery of the usefulness of the messages, integral parts of the ELC learning process, may have been lost owing to the one-on-one nature of the classes. In addition, the implementation of the ELC was not assessed, other than observing classes and giving feedback. Despite this situation, response to the nutrition education classes and the key messages was positive, and caregivers overwhelmingly reported enjoying the intimate character of the classes. Caregivers appeared to value the information given in class and appeared willing to verbalize this knowledge to important people in their lives. In addition, the intervention group’s attitudes about overfeeding appeared to change in a positive direction; the percentage of caregivers who felt that they had to be careful not to feed the child too much increased from 53% to 84% in the intervention group, whereas there was very little change in the control group. Another possibility is that the educational sessions, although convincing, did not adequately provide the specific tools or skills needed to achieve behavioral change. Almost half of caregivers in the intervention group said that the most interesting class topic was the information on ways to calm a fussy infant. Because feeding is often used to calm a fussy infant,12,13 and because caregivers were being instructed not to use food as the first strategy to comfort their infants, the intervention group was provided with a popular video of calming techniques. Originally, the plan was to show a somewhat shortened version of the video “The Happiest Baby on the Block” during each intervention class. However, owing to time constraints, educators instead sent the video home with caregivers, with the instruction to view it at home and to call with any questions. It is possible that, had there been sufficient time to view the video with each class and to respond to questions and reinforce techniques, educators could have addressed this concept more fully and thereby facilitated greater change in feeding practices. Educators did not ask if the videotape was viewed. During the class, early and late satiety cues were demonstrated with a video of puppets, but it might have been more effective to use video of real infants actually exhibiting these cues. Training caregivers to recognize the cues of their own infants may be an even more effective way to transmit both knowledge and skills for responsive feeding. Additionally, the curriculum may be more effective if offered as several shorter class sessions to build the caregivers’ skills and confidence. The timing of the intervention also may not have been ideal. It may be more appropriate to begin education on responsive feeding prenatally, as there are some indications that the unlearning of appetite regulation can begin shortly after birth.17 Extending the intervention past the point when the majority of caregivers would normally have begun to transition to bottle feeds of at least 6 ounces may also have been beneficial.
250
Kavanaugh et al/INTERVENTION TO MODIFY BOTTLE-FEEDING BEHAVIORS AMONG WIC MOTHERS
It is important to note that the sample size achieved was far lower than originally planned, as only about 18% of eligible subjects enrolled and successfully completed the study. The caregivers who attended class may have been more motivated, or had fewer time constraints, than the remainder of the eligible subjects, which may limit the generalizability of the findings.
IMPLICATIONS FOR RESEARCH AND PRACTICE The large increase in formula intake during the study, from about 800 mL per day to more than 1100 mL per day, is much greater than what is observed among breastfed infants during the same age interval8 and suggests that these infants were being overfed by the end of the study. Many caregivers felt that they were already using responsive feeding techniques, so although they agreed with the messages, they may not have felt the need to change their behavior. A certain percentage of “bottle-emptying” is to be expected, in other words, the infant will finish the bottle some of the time, without any prompting. What is unknown is whether infant-driven bottle-emptying can account for the relatively common occurrence of empty bottles observed (50% to 60%). It is likely that caregiver-driven bottleemptying accounts for a considerable proportion of this phenomenon, though it may not be recognized as a factor by the caregivers themselves. If this scenario is correct, what are the barriers to reducing bottle-emptying? In the focus group discussions conducted during the message development phase, the investigators found that caregivers were concerned with preventing their infants from crying too much and encouraging them to sleep through the night. The messages, which were focused on preventing child overweight (a longerterm outcome), may have been less compelling than the more short-term concerns of coping with a fussy infant. Caregivers use feeding as a strategy for soothing their infants,11-13,18,19 thus it may be difficult to change a behavior that they perceive as having an immediate benefit to them. In summary, the results of this study indicate that formula intakes of infants in this population are quite high— probably reflective of overfeeding—and that modifying bottle-feeding behaviors to prevent overfeeding is a challenging task. Further research is needed to understand the attitudes and life circumstances that are constraints to changing these behaviors.
ACKNOWLEDGMENT This study was supported by the Economic Research Program (ERS) Small Grants Program and the Food Stamp
Nutrition Education Program (FSNEP) of California. The authors wish to thank Dr. Harvey Karp for his assistance with this project.
REFERENCES 1. Ong KKL, Ahmed ML, Emmett PM, Preece MA, Dunger DB. Association between postnatal catch-up growth and obesity in childhood: prospective cohort study. BMJ. 2000;320:967-971. 2. Stettler N, Zemel BS, Kumanyika S, Stallings VA. Infant weight gain and childhood overweight status in a multicenter cohort study. Pediatrics. 2002;109:194-199. 3. Cameron N, Pettifor J, De Wet T, Norris S. The relationship of rapid weight gain in infancy to obesity and skeletal maturity in childhood. Obes Res. 2003;11:457-460. 4. Dewey KG. Growth characteristics of breast-fed compared to formulafed infants. Biol Neonate. 1998;74:94-105. 5. Dewey KG, Heinig MJ, Nommsen LA, Peerson JM, Lonnerdal B. Breastfed infants are leaner than formula-fed infants at 1 year of age: the DARLING study. Am J Clin Nutr. 1993;57:140-145. 6. Dewey KG, Peerson JM, Brown KH, et al. Growth of breast-fed infants deviates from current reference data: a pooled analysis of US, Canadian, and European data sets. Pediatrics. 1995;96:495-503. 7. Dewey KG. Is breastfeeding protective against child obesity? J Hum Lact. 2003;19:9-18. 8. Dewey KG, Nommsen-Rivers L, Lonnerdal B. Plasma insulin and insulin-releasing amino acid (IRAA) concentrations are higher in formula-fed than in breastfed infants at 5 mo of age [meeting abstract]. FASEB J. 2004;18(suppl):A843. 9. Dewey KG, Lonnerdal B. Infant self-regulation of breast milk intake. Acta Paediatr Scand. 1986;75:893-898. 10. Baughcum AE, Burklow KA, Deeks CM, Powers SW, Whitaker RC. Maternal feeding practices and childhood obesity. Arch Pediatr Adolesc Med. 1998;152:1010-1014. 11. Boyington JA, Johnson AA. Maternal perception of body size as a determinant of infant adiposity in an African-American community. J Natl Med Assoc. 2004;96:351-362. 12. Savage SA, Reilly JJ, Edwards CA, Durmin JV. Weaning practices in the Glasgow Longitudinal Infant Growth Study. Arch Dis Child. 1998;79:153-156. 13. Bentley M, Gavin L, Black MM, Teti L. Infant feeding practices of low-income, African-American, adolescent mothers: an ecological, multigenerational perspective. Soc Sci Med. 1999;49:1085-1100. 14. Baker CA, Jensen PJ, Kolb DA. Conversational Learning: An Experiential Approach to Knowledge Creation. Westport, Conn: Quorum Books; 2002. 15. Karp H. The Happiest Baby on the Block: The New Way to Calm Crying and Help Your Newborn Baby Sleep Longer. New York, NY: Bantam Dell; 2002. 16. de Onis M, Onyango AW, Van den Broeck J, Chumlea WC, Martorell R for the WHO Multicentre Growth Reference Study Group. Measurement and standardization protocols for anthropometry used in the construction of a new and international growth reference. Food Nutr Bull. 2004;25:S27-S36. 17. Dollberg S, Lahav S, Mimouni FB. A comparison of intakes of breast-fed and bottle-fed infants during the first two days of life. J Am Coll Nutr. 2001;20:209-211. 18. Hannon PR, Willis SK, Bishop-Townsend V, et al. African-American and Latina adolescent mothers infant feeding decisions and breastfeeding practices: a qualitative study. J Adol Health. 2000;26:399-407. 19. Bronner YL, Gross SM, Caufield L, et al. Early introduction of solid foods among urban African-American participants in WIC. J Am Diet Assoc. 1999;99:457-461.