Associations between technology use, responsive feeding, and child temperament among prior prenatal intervention participants

Abstract

Objectives:

The primary objective was to examine associations between mothers’ television and mobile device (TV/MD) use and responsive feeding during an observed mother-toddler mealtime interaction. The secondary objective was to assess whether dimensions of child temperament were associated with mothers’ TV/MD use.

Methods:

Participants from a prenatal lifestyle intervention trial to prevent excess gestational weight gain among women with overweight and obesity (N = 77) were observed during a dinner time meal when their children were aged 19.4 ± 0.9 months. Trained video coders used the Responsiveness to Child Feeding Cues Scale to rate child strength of early/subtle, positive active, and negative active satiation cues and maternal responsiveness to these cues. Coders also recorded mothers’ use of TV/MD. Child temperament was reported by mothers via the Infant Behavior Questionnaire – Revised Very Short Form.

Results:

Twelve percent (n = 9) of mothers used TV/MD during the mealtime interaction. Children whose mothers used TV/MD exhibited stronger early/subtle cues (4.1 ± 0.4) compared to children whose mothers did not use TV/MD (3.4 ± 0.2; p = 0.04). Mothers who used TV/MD exhibited significantly lower responsiveness to child satiation cues (2.0 ± 0.4) than those who did not use TV/MD (3.4 ± 0.2; p = 0.001). Greater child temperamental negative affectivity was associated with greater likelihood of maternal TV/MD use (OR = 4.80, 95% CI = 1.21, 19.03).

Conclusion:

Mothers’ TV/MD use was associated with greater child temperamental negative affectivity and lower responsiveness to child cues.

Responsive feeding, characterized by caregiver sensitivity and contingent responsiveness to child feeding cues,¹ is hypothesized to support children’s developing abilities to self-regulate intake in response to physiological needs (e.g., feelings of hunger and fullness) rather than external demands (e.g., the amount of food available).² Recent obesity prevention efforts support the importance of responsive feeding and parenting for promoting healthy eating behaviors and weight gain trajectories during early childhood.^{3, 4} Despite widespread recognition of the benefits of responsive feeding, emerging research suggests individual and contextual factors may impair parents’ responsiveness to their children’s cues during mealtimes.

In particular, today’s parents are surrounded by various forms of technology, which are powerful contextual factors affecting family interactions.⁵ Many parents report they frequently watch television (TV) or engage with mobile devices (MD) in the presence of their children,^6–8 especially when they need to relieve stress related to dealing with difficult parent-child interactions.^{9, 10} Previous research examining parents’ use of TV or MDs during mealtime interactions has predominantly focused on parents of young infants or school-age children, and findings illustrate parents’ engagement with technology is associated with poorer parent sensitivity and responsiveness to child cues and lower quality parent-child interactions.^11–14 However, there is also evidence for bidirectional associations between parent technology use and children’s temperament and behavior: parents are more likely to use technology when their children have greater levels of negative affectivity (a dimension of temperament)¹⁵ and greater externalizing behavior (e.g., tantrums),¹⁶ but greater parent technology use may also elicit more negative affectivity and externalizing behavior from children over time.^16–18

Few studies have examined the issue of parent technology use during toddler mealtimes, which is an important period for the development of eating behaviors as children are becoming more autonomous during feeding, but also more challenging to feed due to increases in neophobia and food selectivity.¹⁹ In addition, examination of associations between technology use, responsive feeding, and child temperament may be particularly relevant for mothers with obesity given previous research suggesting mothers with obesity are less involved during their young children’s mealtimes.^20–23 Research is needed to understand associations between mothers’ technology use and responsive feeding within toddler feeding contexts, as well as whether child characteristics, such as temperament, are related to technology use. To this end, the present study aimed to: 1) examine associations between mothers’ TV/MD use and responsive feeding during an observed mealtime interaction and 2) assess whether dimensions of child temperament were associated with mothers’ TV/MD use.

METHODS

Participants

Participants included women (N = 77) who previously participated in a randomized clinical trial to test the effectiveness of a multisite prenatal lifestyle intervention to prevent excess gestational weight gain (the Healthy Beginnings/Comienzos Saludables study) and their toddlers. Study sites included California Polytechnic State University, California and Miriam Hospital, Rhode Island. Details about study recruitment, methods, and primary findings can be found elsewhere.²⁴ In brief, eligibility criteria included: 1) pregnant women between 9–16 weeks gestation, 2) singleton pregnancy, 3) English- or Spanish-speaking, 4) 18 years of age or older, and 5) Body Mass Index (BMI) ≥ 25. Exclusion criteria included: 1) major physical or mental health problems, 2) history of bariatric surgery, 3) contraindications to aerobic exercise, 4) loss of contact during initial screening, and 5) other less frequent health risks and conditions. The Healthy Beginnings/Comienzos Saludables intervention was limited to the prenatal period and promoted healthy eating and activity patterns through behavioral modification strategies. One session focused on strategies to reduce sedentary screen time but the topic of “distracted feeding” was not addressed. The intervention reduced excess gestational weight gain,²⁴ but did not affect birth weight or mothers’ or children’s weight change between birth and 12 or 36 months postpartum.^{25, 26}

For the present follow-up study (the Mealtime for Toddlers study), all mothers who participated in the Healthy Beginnings/Comienzos Saludables study and whose children were 18 months of age were invited to participate in additional assessments. Participants were told that the focus of the study was to explore mother-toddler feeding interactions and risks for childhood obesity. Additional exclusion criteria were child diagnoses that impacted feeding, feeding interactions, and/or growth, including: 1) chronic health conditions, 2) significant food allergies and/or dietary restrictions, 3) psychiatric conditions, or 4) developmental delays. All procedures were approved by the Institutional Review Boards at the Miriam Hospital and California Polytechnic State University; the study was determined to be exempt by the Temple University Institutional Review Board due to no participant enrollment or data collection being done at this site and because data transferred to this site was considered de-identified. All participants provided written informed consent prior to participation.

Procedures

Mealtime Observation at 18 months to Assess Maternal Responsiveness to Child Satiation Cues.

Mother-child interactions were video recorded during a dinner-time meal in the family’s home when children were approximately 18 months old. The mother determined the foods offered at the meal and where the meal took place (e.g., at the dining room table, in the living room). Because the research team was interested in observing mother-child interactions, they requested that other family members not be present at the meal. Once the mother and child were situated in the meal location and ready to eat, a research assistant set up a digital recorder in front of the dyad, then stepped into a different room to minimize influence on the dyad’s interactions or eating behaviors. The mother was instructed to feed her child as she typically would at home and to let the research assistant know when the meal was over. Recording commenced at the start of the meal and stopped when the mother indicated the meal was over. After the mother indicated the meal was over, she rated how similar the dinner meal was to other dinner meals at home on a 9-point Likert-type scale (1=not at all similar to other meals to 9=very similar to other meals).

Trained coders masked to study hypotheses (n = 3 coders for videos in English; n = 2 for videos in Spanish) rated videos using the Responsiveness to Child Feeding Cues Scale (RCFCS), a reliable and valid observational coding scheme to assess maternal responsiveness to child cues.²⁷ This scale allows for coding of maternal responsiveness to child hunger and satiation cues, but the set-up of mealtime observations did not allow for reliable coding of child hunger cues and maternal responsiveness to these cues. Coders rated the strength of the satiation cues exhibited by children during the meal including early/subtle satiation cues (e.g., slows pace of feeding) and more potent active satiation cues, which could be positive (e.g., pushes food away) or negative (e.g., fussing and whining) on a 5-point Likert-type scale (1=very weak to 5=very strong). Coders also rated mothers’ responsiveness to these satiation cues using a 5-point Likert-type scale (1=highly unresponsive to 5=highly responsive). Whether the mother watched TV or used MDs (e.g., texted on a mobile phone) during the meal (referred to hereafter as TV/MD use) was also recorded. Mothers were categorized as “TV/MD used” if they watched TV or used a MD at any point during the mealtime interaction or “No TV/MD used” if they did not watch TV or use a MD. Intercoder reliability was determined by cross-coding of 20% of videos. Intraclass correlation coefficients (ICCs) were used to determine reliability, which was deemed adequate (ICC = 0.58 for videos in English and 0.60 for videos in Spanish).

Child Temperament.

Mothers completed the Infant Behavior Questionnaire - Revised Very Short Form (IBQ-RVS) when their children were 18 months old.²⁸ The IBQ-RVS is a 37-item instrument that measures three dimensions of child temperament: Positive Affectivity/Surgency, Negative Affectivity, and Orienting/Regulatory Capacity. Questionnaire items were scored on a Likert scale of 1 to 7, with higher scores representing greater Positive Affectivity/Surgency, Negative Affectivity, or Orienting/Regulatory Capacity. The IBQ-RVS has demonstrated validity and good internal consistency (Positive Affectivity/Surgency, α = 0.77; Negative Affectivity, α = 0.78; Orienting/Regulatory Capacity, α = 0.91) in diverse populations of mothers of children younger than 3 months and up to 3 years old.²⁸ In the present study, subscales also demonstrated acceptable internal consistency (Positive Affectivity/Surgency, α = 0.71; Negative Affectivity, α = 0.79; Orienting/Regulatory Capacity, α = 0.77).

Other Maternal, Child, and Family Characteristics.

At study entry, mothers reported basic demographic information, including age, annual family income level, education level, race, ethnicity, and parity. At the 12-month postpartum assessment, mothers reported the number of hours per week they spent sitting at home watching TV/VCR/DVDs. Given the wide range of responses provided, responses were dichotomized via median split into 0–5 hours per week versus 6+ hours per week. Weight and height for both mothers and children were measured in duplicate and while in light clothing and no shoes. Mothers’ body mass index scores (BMI) were calculated using the standard equation BMI=kg/m². Children’s age and biological sex-specific BMI z-score (BMIz) were calculated using World Health Organization (WHO) growth standards.²⁹

Statistical Analysis

All analyses were conducted using SAS v.9.4 (SAS Institute Inc., North Carolina, USA). All dyads (N = 77) had complete data for mealtime behaviors and other characteristics. Three mothers did not complete the IBQ-RVS; thus, the sample size for analyses with child temperament was n = 74. Descriptive statistics were calculated to summarize sample characteristics; Fisher’s exact test and general linear models, which have advantages over other approaches when working with small sample sizes,^{30, 31} were used to examine associations between mothers’ TV/MD use and sample characteristics. General linear models were used to examine associations between mothers’ TV/MD use, child strength of satiation cues, and maternal responsiveness to satiation cues, as well as associations between these variables and dimensions of child temperament. Logistic regression with Firth’s Penalized Likelihood to reduce bias caused by sample size³² was used to examine whether dimensions of child temperament were related to mothers’ TV/MD use. All models controlled for treatment group and study site; maternal age, education level, ethnicity, and parity; family income level; and child sex and age. Statistical significance was defined as p < 0.05.

RESULTS

Sample Characteristics

Table 1 presents sample characteristics. The average meal length was 18.9 ± 5.6 minutes (range = 8.3 – 34.9 minutes). On average, mothers reported the dinner meal was similar to other dinner meals at home (6.8 ± 2.2). Twelve percent (n = 9) of mothers used TV/MD during the mealtime observation. Sample characteristics did not significantly differ for mothers who did versus did not use TV/MD.

Table 1.

Sample Characteristics for the Total Sample and by Mothers’ TV/MD Use during an Observed Mother-Child Mealtime Interaction

	Total Sample		No TV/MD Use		TV/MD Use		p-valuea
	N = 77		88.3%, n = 68		11.7%, n = 9
	% (n) or mean (SD)	Range	% (n) or mean (SD)	Range	% (n) or mean (SD)	Range
Group, % (n)							0.16
Intervention	48.1 (37)		51.5 (35)		22.2 (2)
Standard Care	52.0 (40)		48.5 (33)		77.8 (7)
Site, % (n)							0.49
Rhode Island	42.9 (33)		41.2 (28)		55.6 (5)
California	57.1 (44)		58.8 (40)		44.4 (4)
Mother Characteristics
Age at Study Entry, mean (SD) years	30.4 (4.5)	19.0–41.0	30.5 (4.7)	19.0–41.0	29.7 (3.0)	26.0–34.0	0.59
Body Mass Index, mean (SD) kg/m2	33.9 (6.4)	22.9–47.2	34.1 (6.4)	22.9–47.2	33.1 (6.8)	25.2–47.1	0.66
Annual Family Income Level, % (n)							0.29
<$50,000	42.9 (33)		45.6 (31)		22.2 (2)
$50,000+	57.1 (44)		54.4 (37)		77.8 (7)
Education Level, % (n)							0.61
High School Degree or Less	24.7 (19)		26.5 (18)		11.1 (1)
Some College or College Degree	59.7 (46)		58.8 (40)		66.7 (6)
Graduate Degree	15.6 (12)		14.7 (10)		22.2 (2)
Ethnicity, % (n)							0.99
Not Hispanic	59.7 (46)		60.3 (41)		55.6 (5)
Hispanic	40.3 (31)		39.7 (27)		44.4 (4)
Race, % (n)							0.57
White	59.7 (46)		54.6 (42)		5.2 (4)
Black	5.2 (4)		5.2 (4)		0 (0)
Asian	2.6 (2)		2.6 (2)		0 (0)
Mixed	31.2 (24)		24.7 (19)		6.5 (5)
Not Reported	1.3 (1)		1.3 (1)		0 (0)
Parity, % (n)							0.68
Primiparous	20.8 (16)		22.1 (15)		11.1 (1)
Multiparous	79.2 (61)		77.9 (53)		88.9 (8)
Average hours per week spent watching TV, % (n)							0.99
0–5 hours per week	48.1 (37)		48.5 (33)		44.4 (4)
6+ hours per week	39.0 (30)		38.2 (26)		44.4 (4)
Not Reported	13.0 (10)		13.2 (9)		11.1 (1)
Child Characteristics
Sex, % (n)							0.72
Male	42.9 (33)		44.1 (30)		33.3 (3)
Female	57.1 (44)		55.9 (38)		66.7 (6)
Age at Assessment, mean (SD) months	19.4 (0.9)	17.9–21.2	19.5 (0.9)	17.9–21.2	19.0 (0.6)	18.0–20.2	0.13
BMI z-score	0.7 (1.0)	−3.5–2.7	0.7 (1.1)	−3.5–2.7	0.4 (0.7)	−0.4–2.21	0.30

BMI, Body Mass Index; SD, Standard Deviation, TV/MD, Television and/or Mobile Device

^aFor Fisher’s Exact Tests and general linear models testing for differences between No TV/MD Use versus TV/MD Use groups

TV/MD Use and Mealtime Behaviors

Children of mothers who used TV/MD exhibited stronger early/subtle satiation cues (4.1 ± 0.4) compared to children of mothers who did not use TV/MD (3.4 ± 0.2; p = 0.04; Figure 1, Panel A). Strength of positive active cues was not significantly different for children of mothers who used TV/MD (3.7 ± 0.4) versus did not use TV/MD (3.9 ± 0.2; p = 0.68; Figure 1, Panel B). Strength of negative active cues was also not significantly different for children of mothers who used TV/MD (3.3 ± 0.6) versus children of mothers who did not use TV/MD (2.3 ± 0.3; p = 0.08; Figure 1, Panel C). Although there were observed differences in strength of early/subtle satiation cues exhibited by their children, mothers who used TV/MD exhibited significantly lower responsiveness to child satiation cues (2.0 ± 0.4) than mothers who did not use TV/MD (3.4 ± 0.2; p = 0.001) (Figure 1, Panel D).

Figure 1. Associations between mothers’ TV/MD use during an observed meal, child strength of satiation cues, and maternal responsiveness to child cues (N = 77).

Panels present adjusted mean differences between mothers who did not versus did use TV or MDs during an observed maternal-child mealtime interaction. Groups were significantly different for child strength of early/subtle satiation cues (Panel A, p = 0.04), but not for strength of positive active cues (Panel B, p = 0.68) or negative active cues (Panel C, p = 0.08). Groups were also significantly different for maternal responsiveness to child cues (Panel D, p = 0.001). All models controlled for treatment group and study site; maternal age, education level, ethnicity, and parity; family income level; and child sex and age.

Child Temperament

Greater maternal-reported child negative affectivity was associated with significantly greater likelihood of maternal TV/MD use (OR = 4.80, 95% CI = 1.21, 19.03). Reported child positive affectivity/surgency (OR = 0.19, 95% CI = 0.02, 1.91) and orienting/regulatory capacity (OR = 2.13, 95% CI = 0.42, 10.89) were not significantly associated with maternal TV/MD use. Dimensions of maternal-reported child temperament were not significantly associated with maternal responsiveness nor with child strength of early/subtle satiation cues, positive active cues, or negative active cues (data not shown).

DISCUSSION

The present study is among the first to examine associations between TV/MD use, responsive feeding, and child temperament within a diverse sample of mothers and toddlers. Within our sample, twelve percent of mothers engaged with TV/MD during the observed mealtime interaction. These mothers exhibited significantly lower responsiveness to child satiation cues, despite having children who displayed stronger early/subtle cues, compared to mothers who did not use TV/MD. In addition, greater levels of child temperamental negative affectivity were associated with greater likelihood of mothers’ TV/MD use.

The present findings align with previous studies focused on other age groups. Studies of mothers of young infants illustrated mothers distracted by technology were less sensitive to infant cues and engaged their infants less during feeding.^{11, 14} Likewise, observations of parents engaged in mealtimes with older children illustrated parents who used MDs were less likely to respond to their children’s bids for attention and, when they did respond, they did so in a manner that was slower and less attentive than parents who were not using a MD.^{13, 33} Taken together, these findings illustrate parent technology use during mealtimes is associated with lower levels of responsive feeding and lower quality interactions between parents and children, even during the toddler years.

Children of mothers who used TV/MD exhibited greater strength of early/subtle satiation cues compared to children of mothers who did not use TV/MD. This finding aligns with observational research of family mealtimes that illustrates older children respond to parent MD use by amplifying their bids for attention.¹³ Similarly, experimental research with 7- to 24-month-olds illustrates that infants increase their frequency of social bids for their mothers’ attention when their mothers disconnect from face-to-face play interactions to attend to a MD.¹⁷ Thus, one possible interpretation of these associations is that children of distracted mothers “speak up” by increasing the strength and frequency of certain types of cues and bids for attention. However, we did not find associations between maternal TV/MD use and child strength of positive or negative active satiation cues, which is inconsistent with this interpretation. Given the cross-sectional and short-term nature of the present study and most other studies on this topic, further experimental and longitudinal research is needed to better understand potential associations between maternal TV/MD use, children’s behavior, and communication during feeding.

Maternal TV/MD use was associated with greater levels of child temperamental negative affectivity, which is consistent with previous research illustrating that parents are more likely to engage with technology within family contexts when they perceive their children to be more challenging.^{9, 15} Parents report using technology to cope with difficult aspects of parenting, such as feelings of boredom or stressful interactions with children.⁹ Further, parents who feel stressed by difficult interactions with their child report using their MDs as a means to withdraw and cope with this stress, which may then exacerbate their children’s difficult behaviors over time.^{16, 34} Thus, one possible interpretation of the findings of the present study is that mothers who perceived their children to be more challenging (as indicated by greater levels of reported negative affectivity) were more likely to use TV/MD during mealtimes as a means to cope.

However, it is also possible that mothers’ TV/MD use elicits greater levels of negative affectivity from children. Previous experimental research with 7–24-month-olds illustrates that when mothers use MDs during mother-infant interactions, infants increase their display of negative affect and decrease their display of positive affect.^{17, 18} Observational research suggests greater levels of problematic technology use (i.e., technology use that interferes with family interactions) for mothers predicts greater internalizing (e.g., withdrawal) and externalizing (e.g., tantrums) behaviors for children over time.³⁵ Thus, greater child negative affectivity may be a result of mothers’ habitual use of TV/MD during feeding and other interactions, perhaps influenced by lower maternal responsiveness during these interactions.

Our sample was unique in that mothers had overweight or obesity during early pregnancy and participated in a lifestyle intervention to prevent excessive gestational weight gain. Although 68% of the sample still had overweight or obesity at the time of the mealtime observation (data not shown), the study’s small sample size did not allow us to examine whether mothers’ or children’s weight status modified noted associations. Some, albeit limited, research comparing the feeding practices and styles of mothers with healthy weight versus overweight or obesity suggests that mothers with overweight or obesity report less involvement during children’s mealtimes. Specifically, mothers with overweight or obesity were more likely to report giving their children control around eating and less likely to encourage dietary balance and variety, teach their child about nutrition, and model healthy eating.^20–23 However, further research is needed to better understand whether maternal TV/MD use, or associations between maternal TV/MD use, responsive feeding, and child temperament, is associated with mothers’ or children’s weight status.

Given the cross-sectional nature of the present study, further longitudinal research is also needed to better understand potential bidirectional effects between maternal TV/MD use, responsive feeding, and child temperament. Future research should also aim to address other study limitations. The present study was limited by the small sample size, warranting additional research in larger, more diverse samples. Although our use of observational coding of mother-child mealtime interactions was a strength, the ICC of observational scores was adequate but low, reflecting reduced reliability of our observational measures. In addition, maternal TV/MD use was coded as a binary variable (No TV/MD Use vs TV/MD Use); further research with more nuanced assessment of the nature and extent of TV/MD is needed. We conducted feeding observations of one meal within participants’ homes to promote ecological validity, but participants’ behaviors may have been biased because mothers and children knew they were being filmed and/or one meal may not have been representative of all mealtime interactions. In addition, child temperament was based on maternal report, which provides valuable information regarding mothers’ perceptions of child behavior but may also be susceptible to reporting bias. Further research that includes observational measures of child temperament may provide additional insights regarding associations between maternal TV/MD use and temperament. Despite these limitations, the present study is among the first to examine the issue of TV/MD use during toddler mealtimes and provides initial insights that can inform the design and focus of future studies of this population.

In conclusion, responsive feeding supports healthy mealtime interactions between parents and children during the toddler years. The increasing presence of distracting technologies, such as TV and MDs, within the home has led to increasing concern that these technologies may interfere with parents’ abilities to be attentive and responsive to their children’s cues and needs. Findings from the present study illustrate associations between mothers’ technology use, responsive feeding, and children’s behaviors within a diverse sample of mothers with overweight or obesity during early pregnancy and who participated in a prenatal lifestyle intervention to prevent excess gestational weight gain, but further longitudinal research is needed to understand long-term implications of these findings for child weight and developmental outcomes.