Longitudinal Associations Between Family Meal Quality and Quantity: Does one Matter More for Child, Parent, and Family Health and Well-being or are they Synergistic?

Abstract

This study aimed to examine longitudinal associations between family meal quantity (i.e., frequency) and quality (i.e., meal healthfulness and interpersonal quality) and child, parent, and family health and well-being and whether there was a synergistic effect between family meal quantity and quality. Children ages 5–9 and their parents from six racial/ethnic groups participated in this longitudinal cohort study. Regression models examined family meal quantity, interpersonal quality, and nutritional quality at baseline and interactions between quantity and quality, with changes in child, parent, and family health outcomes from baseline to 18-month follow-up. Higher family meal quantity predicted reduced obesity prevalence, improved diet quality and less food fussiness, food responsiveness, and conduct problems among children at follow-up. Higher family meal quality predicted improved diet quality, lower emotional problems, and fewer peer relationship problems among children, improved diet quality for parents, and less family chaos at follow-up. An interaction between family meal quantity and quality was found for child peer relationship problems. Family meal quantity and quality were independently important for child health and well-being and for some parent and family health outcomes. Clinicians may want to emphasize the importance of both family meal quantity and quality during well-child visits, as these longitudinal findings suggest potential benefits for the entire family.

1. INTRODUCTION

Over two decades of cross-sectional^1–16 and longitudinal research^17–25 have shown higher family meal quantity (i.e., frequency) is protective for adolescent health including higher diet quality, less unhealthy weight-control behaviors, better psychosocial health, and reduced risk for adolescent obesity–although findings with weight status have been mixed.^19,20,26,27 These protective associations between family meal quantity and adolescent health and well-being outcomes have also held across race/ethnicity, age, sex, and income.^{13,14,28–30} However, few studies have been conducted with children to examine whether similar protective associations exist between family meal quantity and child health and well-being. The few studies that have been conducted with children have found mixed results with some showing associations with lower weight status and some finding no associations. ^20,23,26,27 In addition, the majority of research conducted with children has been cross-sectional and included less racially/ethnically, socioeconomically, and educationally diverse populations. Furthermore, some cross-sectional studies have shown that higher family meal quantity is associated with better diet quality for adults,^28,31–33 suggesting that family meals may potentially benefit the entire family. Thus, more research is needed to confirm associations between family meal quantity and children and parent health and well-being in longitudinal studies.

Less is known about family meal quality, in terms of both diet quality (i.e., healthfulness of food served at the meal) and interpersonal quality (i.e., emotional atmosphere) and associations with child, parent, and family health and well-being. Some cross-sectional^9,34–39 and longitudinal⁴⁰ studies have found that higher diet quality of foods served at family meals (e.g., fruits/vegetables, whole grains) and quality of interpersonal interactions during family meals (e.g., non-controlling parent feeding practices, communication) were associated with lower weight status, higher diet quality, and emotional well-being in children^9,35,41 and adolescents.^9,36,40 However, the majority of these studies were cross-sectional and only explored one aspect of family meal quality, either healthfulness of the meal or interpersonal interactions. Additionally, most prior research has not included family health and well-being outcomes. More research is needed that simultaneously examines both meal healthfulness and interpersonal interactions at family meals with child, parent, and family health and well-being outcomes.

Furthermore, it is unclear whether family meal quantity matters more than quality, or if these two factors have a synergistic influence on child, parent, and family health and well-being. Thus, the two-fold purpose of this study is to build upon and expand prior research by (1) examining longitudinal associations between family meal quantity and quality and child, parent, and family-level health and well-being and (2) examining whether family meal quantity and quality act synergistically in a racially/ethnically diverse and immigrant/refugee sample.

Family Systems Theory (FST)^42–44 guides the current study. According to FST, family functioning—via patterns and processes families engage in daily—is thought to shape children’s health and well-being. For example, establishing routines/patterns, such as having regular family meals creates an environment of safety, security, and predictability, which is expected to be protective for child health and well-being. In addition, FST suggests that both the routine itself (i.e., family meal quantity) and what occurs during the routine (i.e., family meal quality) influences child health and well-being outcomes. Furthermore, one FST tenet is that the “whole is greater than the sum of its parts”. For example, family meal quantity and quality may have a synergistic influence on health and well-being outcomes compared to the effect of either one of them individually. FST also suggests that families who engage in routines together are more likely to see benefits for all family members. Thus, we hypothesize that family meal quantity and quality (i.e., diet quality and interpersonal quality) will each predict longitudinal improvements in child, parent, and family health and well-being, and that there will be a synergistic effect between family meal quantity and quality.

2. MATERIAL AND METHODS

Data for the current study are from the Family Matters study,⁴⁵ a mixed-methods study of children ages 5–9 years old (n=1307) and their families from low-income, racially/ethnically diverse, and immigrant/refugee households. At two time points 18-months apart (2016 and 2018), parents completed an online survey. At follow-up, 96% of the Family Matters sample was retained (n=1259).

2.1. Recruitment.

Parent/child dyads were recruited from primary care clinics in Minneapolis/St. Paul. Families were eligible to participate if: (1) the child was 5–9 years old; (2) the person completing the survey was the primary guardian of the child; (3) the child lived with the parent/guardian more than 50% of the time; (4) the child was from one of the following racial/ethnic backgrounds: African American, Hispanic/Latino, Hmong, Native American, Somali, Ethiopian, or White; and (5) the child had a body mass index greater than the 5^th percentile. The study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the Institutional Review Board Human Subjects Committee at the University of Minnesota (IRB #1107S02666). Written informed consent and assent was obtained from all study participants before participating in the study. All study materials were translated into Hmong, Somali and Spanish. In-depth details about the study are published elsewhere.⁴⁵

2.2. Survey Development:

Development of the online survey followed best practice,^7–9 including: (1) using established valid and reliable measures; (2) using survey psychometrics from our pilot study of 150 families⁴⁶ to identify which survey items to keep, cut, or alter; and (3) conducting a test-retest reliability sub-study with 125 participants (~20 per race/ethnicity) from the current sample who took the survey twice within two weeks. This sub-study estimated intraclass correlation coefficients (ICC) from mixed models with participant-level random intercepts to capture the degree of agreement between measurements. Agreement was high (ICC>0.8) for more objective questions (e.g., receives Supplemental Nutrition Assistance Program benefits, relationship status, height, household size), moderate (ICC range: >0.6 - <0.8) for questions about most food-related attitudes and behaviors (e.g., meal frequency, food shopping and preparation attitudes and behaviors, child eating behaviors), and low (ICC<0.6) for questions that are expected to vary over time (e.g., stress, dietary intake, home food availability).

2.3. Sample demographics.

The analytic sample for the present study included 1,246 children and 1,077 caregivers (155 families have more than one child enrolled in the study) with sufficient data to examine at least one outcome of interest at follow-up. These families were well distributed across the six racial/ethnic groups recruited (23.0% Black or African American, 17.0% Hispanic or Latino, 14.8% Hmong, 14.6% Native American, 8.7% Somali or Ethiopian, and 21.9% White). Children in the analytic sample were, on average, 6.9 years old at baseline (SD=1.5) and 49.3% were girls. Caregivers were, on average, 35.7 years old at baseline (SD=7.7), 89.8% were mothers, and 37.5% had not attended school beyond a high school diploma or the equivalent.

2.4. Measures

All measures are described in Table 1, in addition to how variables were operationalized.^{5,35,45,47–61}

Table 1:

Description of Predictor, Outcome, and Control Variables used in Analyses.

Construct	Question	Response Options	How Variable was Operationalized
Predictor Variables
Family Meal Quantity	Parents were asked the following question via an online survey at both time points: “During the past week, how many times were most members of your family sitting and eating a family meal together?”5	Never, 1–2 times, 3–4 times, 5–6 times, 7 times, and more than 7 times	A continuous variable was created using the midpoint of each response range and 8 for the highest category.
Family Meal Quality (i.e., meal healthfulness and interpersonal atmosphere)
Nutritional quality of foods served at dinner	Parents were asked to report on the frequency that the following foods are served at a typical family dinner at both time points: dark-green vegetables; other vegetables; fried vegetables; fruit; 100% fruit juice; dairy products; whole grains; refined grains; plant proteins; sugar sweetened drinks; salty snacks; baked goods; and candy.45	Never or rarely, Sometimes, Usually, Always	The following 10 food categories were created: dark-green vegetables/legumes; other vegetables; total fruit (including 100% juice); dairy; plant proteins; refined grains; whole grains; salty snacks; added sugars; saturated fats.35 Each category was assigned a value of 0–3 (0=Never; 3=Always), with the starred categories reverse coded (i.e., families received points for not serving the food category). Points were then summed across categories, for a score ranging from 0–30.
Interpersonal quality of interactions at dinner	Parents were asked to report which of the following things typically occur during a family meal in your home? (check all that apply)35,45	Conversations, Watching TV, TV on in background, Playing a video game, Using a cell phone, Using a tablet, Using a computer, Reading/looking at a book, Listening to headphones, None of the above.	Responses were dichotomized such that family meals involving no conversation or involving any media distractions were considered representative of less positive family meal interpersonal quality, and family meals involving conversation without media distractions were considered representative of more positive family meal interpersonal quality.
Outcome Variables
Child and Parent Weight
Child Weight	Baseline child height and weight from electronic medical records (EMR) were provided in recruitment data. Child BMI percentiles were then calculated from electronic medical records according to CDC guidelines.49 At 18-month follow-up, EMR were used again for child height and weight. For those children without a height and weight in EMR because of moving clinics, out of state, or not attending a clinic visit recently, the following strategies were used to collect updated data (in the following order of operations): (1) After Visit Summary (n=209); (2) Height/weight taken by study staff or other provider (e.g., school nurse) (n=65); (3) Study-provided objective measuring packet (n=262); (4) Parent self-report of child’s height/weight (n=154). Child BMI percentiles were then calculated from electronic medical records according to CDC guidelines.49	Child BMI percentiles were calculated using height, weight, sex, and birthdate.49	Children with a BMI %ile ≥ 95th were classified as having obesity.
Parent Weight	During the online survey at baseline and 18-month follow up, parents were asked to self-report their height and weight.	Parent BMI was calculated using the CDC’s online BMI calculator.50	Adults with a BMI ≥ 30 kg/m2 were classified as having obesity
Child and Parent Diet Quality
Child/Parent Diet Quality	Parents were asked to report on their child’s and their own frequency of intake of 15 food categories (e.g., Dark-Green vegetables, Fruit, Refined Grains) over the past month using an adapted version of the Children’s Eating Habits Questionnaire52,53 via an online survey at both time points.	Response options for each category were: Never/Rarely, 1–3 times per month, 1–2 times per week, 3–4 times per week, 5–6 times per week, 1+ times per day.	Using the Healthy Eating Index-201554 as a guide, staff dietitians created an overall dietary intake quality score (for children and for parents) ranging from 0–85, where higher scores indicate higher diet quality. Details regarding the dietary quality score have been published elsewhere.55
Child Eating Behaviors	Children’s eating behaviors were assessed via an online survey at both time points using nine items from the Children’s Eating Behavior Questionnaire (CEBQ).51 • Food responsiveness items: always asking for food, would eat most of the time, would eat too much • Satiety responsiveness items: gets full before their meal is finished, eats slowly, finishes their meal quickly [reverse coded] • Food fussiness items: refuses new foods at first, enjoys a wide variety of foods [reverse coded], decides they do not like a food without tasting it	Responses were recorded on a five-point Likert-type scale with response options ranging from 1 = Never to 5 = Always.	Responses were summed for each subscale; subscale scores each ranged from 3–15, with higher scores representing greater endorsement of the respective eating style.
Child and Parent Emotional and Social Well-Being
Child Emotional Well-Being	Parents reported their children’s emotional problems, conduct problems, hyperactivity, peer relationship problems, and prosocial behavior using the 25-item Strengths and Difficulties Questionnaire (SDQ)56,57 via an online survey at both time points.	0 = Not true, 1 = Somewhat true, and 2 = Certainly true	Items were reverse coded, and responses were summed for each subscale; subscale scores each ranged from 0–10, with higher scores representing greater endorsement of the respective behaviors.
Psychological distress	Parents indicated the frequency with which they had experienced symptoms (felt nervous, hopeless, restless/fidgety, depressed, worthless, that everything was an effort) over the past 30 days using the six-item Kessler Psychological Distress Scale (K6)58 via an online survey at both time points.	0 = None of the time, 1 = A little of the time, 2 = Some of the time, 3 = Most of the time, and 4 = All of the time	Responses were summed; scores ranged from 0–24, with higher scores representing greater psychological distress.
Parent anxiety	Parents indicated the frequency with which they had experienced symptoms (not being able to stop or control worrying, having trouble relaxing, becoming easily annoyed or irritable, feeling afraid as if something awful might happen) over the past two weeks using four items from the seven-item Generalized Anxiety Disorder screening tool (GAD-7)59 via an online survey at both time points.	0 = None of the time, 1 = A little of the time, 1 = Some of the time, 2 = Most of the time, 3 = All of the time.	Responses were summed; scores ranged from 0–12, with higher scores representing greater anxiety symptoms. A score ≥9 was considered indicative of severe anxiety.
Family Well-Being
Family functioning	Family functioning was assessed via an online survey at both time points with the six-item General Family Functioning Sub-scale from the Family Assessment Device (FAD).60	Four-point Likert agreement scale with response options ranging from 0 = Strongly disagree to 3 = Strongly agree.	Appropriate items were reverse coded, and responses were summed. Scores ranged from 0–18, with higher scores representing higher family functioning.
Family chaos	Parents responded to the following items via an online survey at both time points: (1) “We almost always seem to be rushed”, (2) “You can’t hear yourself think in our home”, and (3) “We have regular routines/schedules in our home (e.g., teeth brushing, family meals, bedtimes, chores)” [reverse coded] from the Confusion, Hubbub, and Order Scale (CHAOS).61	Four-point Likert agreement scale with response options ranging from 1 = Strongly disagree to 4 = Strongly agree.	A score ranging from 3–12 was calculated by summing responses, with higher scores indicate higher household chaos.
Control Variables
Household race/ethnicity	Parents were asked via an online survey at both time points to select the race/ethnicity that best characterizes your household (e.g., the foods you cook for your family, the language you speak at home, the holidays you celebrate).	White | Black/African American | Latino | Hmong | Native American | Somali
Parent age at baseline	Parents were asked via an online survey at both time points: “When is your birthday?”
Parent sex	Parents were asked via an online survey at both time points: “Do you identify as…”	Female | Male
Parent educational attainment	Parents were asked via an online survey at both time points: “What is the highest level of education that you have completed?”	Middle school or junior high; Some high school; High school graduate or GED; Vocational, technical, trade or other certification program; Associate degree; Some college; Bachelor’s degree; Graduate or professional degree	Categories were collapsed into: (1) less than high school; (2) high school diploma or equivalent; (3) some college or vocational/associate degree; (4) bachelor’s degree; and (5) graduate/professional degree.
Child age at baseline	Parents were asked via an online survey at both time points: “When is [child]’s birthday?”	Parents selected a calendar date
Child sex	Parents were asked via an online survey at both time points: “Does your child in the study identify as...”	Female | Male

2.5. Statistical Analysis

Generalized estimating equations with an independence correlation structure were used to examine associations between each family meal characteristic and health or well-being outcome while accounting for correlated data within households. For parents with >1 child in the study (who completed separate surveys for each child), only data from the first survey per study wave that each parent completed were used in analyses of parent and family-level health or well-being outcomes. Linear regression and modified Poisson regression⁶² models were conducted for continuous and dichotomous outcomes, respectively. In separate models, each family meal characteristic (i.e., quantity, interpersonal quality, nutritional quality) at baseline was examined as an independent variable in relation to dependent variables representing each health or well-being outcome at follow-up, adjusting for the given outcome at baseline to examine changes over time.⁶³ All models also adjusted for household race/ethnicity, parent age, parent gender, parent educational attainment; models for child and family-level outcomes were additionally adjusted for child age and gender. To examine synergism between family meal quantity and quality, quantity x interpersonal quality and quantity x nutritional quality interaction terms, as well as the appropriate lower-order terms, were added to the models. Interactions were tested on the additive scale, appropriate for testing for synergism.⁶⁴ Significant interactions were interpreted graphically and with simple slope analyses. Analyses were performed in Stata 16.1.

3. RESULTS

3.1. Descriptive Results

At baseline, parents reported an average of 5.5 (SD=2.3) family meals over the past week, the mean score for the nutritional quality of the meals served was 15.5 (SD=4.1; possible range: 0–30), and 46.4% of parents reported positive family meal interpersonal quality. Descriptive statistics for child, parent, and family health and well-being outcomes at baseline and follow-up are reported in Table 2.

Table 2.

Child, Parent, and Family Health and Well-being Outcomes at Baseline and 18-month Follow-Up

	N with Data at Both Time Points	Baseline	Time 1 Follow-Up
		Mean ± SD or % (n)
Child outcomes *
Diet quality (possible range: 0–85)	1,181	57.3 ± 10.4	56.5 ± 10.3
Food responsiveness (possible range: 3–15)	1,176	8.1 ± 2.7	7.9 ± 2.7
Satiety responsiveness (possible range: 3–15)	1,176	9.7 ± 2.2	9.4 ± 2.2
Food fussiness (possible range: 3–15)	1,176	8.5 ± 2.9	8.2 ± 2.9
Obesity	1,219	18.7 (228)	24.9 (303)
Emotional problems (possible range: 0–10)	1,149	1.4 ± 1.9	1.5 ± 2.0
Conduct problems (possible range: 0–10)	1,139	1.9 ± 1.9	1.8 ± 1.9
Hyperactivity (possible range: 0–10)	1,149	3.3 ± 2.4	3.1 ± 2.4
Peer relationship problems (possible range: 0–10)	1,025	2.0 ± 1.7	2.0 ± 1.8
Prosocial behavior (possible range: 0–10)	1,149	8.3 ± 2.1	8.3 ± 2.3
Parent outcomes *
Diet quality (possible range: 0–85)	1,016	56.4 ± 10.6	57.3 ± 10.2
Obesity	1,004	40.5 (407)	40.8 (410)
Psychological distress (possible range: 0–24)	1,000	4.6 ± 5.2	4.3 ± 4.9
Severe anxiety	1,000	5.3 (53)	3.7 (37)
Family outcomes *
Family functioning (possible range: 0–18)	1,000	14.9 ± 2.8	14.9 ± 2.9
Family chaos (possible range: 3–12)	1,000	5.9 ± 1.7	5.9 ± 1.7

SD, standard deviation.

^*Data were available at both baseline and follow-up for at least one child outcome for 1,243 children, and 1,016 parents provided data at both baseline and follow-up for at least one parent or family outcome. N’s differ within child, parent, and family outcomes due to variable-level missingness.

3.2. Associations Between Family Meal Quantity and Child, Parent, and Family Health and Well-Being

3.2.1. Children.

Among children, greater family meal quantity at baseline predicted significantly reduced obesity prevalence, improved diet quality, reduced food responsiveness and food fussiness, and reduced conduct problems 18 months later, adjusting for baseline levels of these outcomes as well as household race/ethnicity, parent age, gender, and educational attainment, and child age and gender (see Table 3). For example, after adjusting for sociodemographic characteristics and obesity at baseline, each additional family meal per week significantly predicted, on average, 4% lower prevalence of obesity at follow-up (p < .05).

Table 3.

Longitudinal Associations Between Family Meal Quantity and Quality and Child Health and Well-being Outcomes

Family Meal Quantity and Quality at Baseline	Eating and Weight Outcomes at Time 1 Follow-Up, Adjusted for Outcome at Baseline
	Diet Quality	Food Responsiveness	Satiety Responsiveness	Food Fussiness	Obesity
	B (95% CI)	B (95% CI)	B (95% CI)	B (95% CI)	PR (95% CI)
Quantity of family meals
Family meal frequency	0.33** (0.12, 0.55)	−0.07** (−0.13, −0.02)	0.00 (−0.05, 0.04)	−0.10*** (−0.15, −0.05)	0.96* (0.93, 0.99)
Quality of family meals
Interpersonal quality	2.36*** (1.36, 3.36)	−0.12 (−0.39, 0.15)	0.03 (−0.19, 0.26)	−0.09 (−0.34, 0.16)	0.99 (0.84, 1.17)
Nutritional quality	0.11 (−0.01, 0.23)	0.00 (−0.03, 0.03)	−0.01 (−0.03, 0.02)	−0.03 (−0.06, 0.00)	1.01 (0.99, 1.03)
Family Meal Quantity and Quality at Baseline	Psychosocial Outcomes at Time 1 Follow-Up, Adjusted for Outcome at Baseline
	Emotional Problems	Conduct Problems	Hyperactivity	Peer Relationship Problems	Prosocial Behavior
	B (95% CI)	B (95% CI)	B (95% CI)	B (95% CI)	B (95% CI)
Quantity of family meals
Family meal frequency	−0.04 (−0.08, 0.01)	−0.04* (−0.08, −0.01)	−0.01 (−0.06, 0.03)	−0.01 (−0.05, 0.03)	−0.02 (−0.03, 0.07)
Quality of family meals
Interpersonal quality	−0.29** (−0.49, −0.09)	−0.06 (−0.24, 0.13)	−0.14 (−0.36, 0.09)	−0.19* (−0.36, −0.01)	0.18 (−0.05, 0.42)
Nutritional quality	−0.01 (−0.03, 0.02)	−0.01 (−0.03, 0.01)	0.00 (−0.03, 0.02)	0.01 (−0.01, 0.04)	0.01 (−0.02, 0.04)

PR, prevalence ratio; CI, confidence interval. Models were adjusted for household race/ethnicity, parent age, parent gender, parent educational attainment, child age, child gender, and outcome of interest at baseline. Example interpretation for a B estimate (corresponding to continuous outcome): after adjusting for sociodemographic characteristics and diet quality at baseline, a more positive family meal interpersonal quality predicted a significantly better diet quality score at follow-up by 2.36 points on average. Example interpretation for a PR estimate (corresponding to dichotomous outcome): after adjusting for sociodemographic characteristics and obesity at baseline, each additional family meal per week significantly predicted, on average, 4% lower prevalence of obesity at follow-up.

^*p < .05,

^**p < .01,

^***p < .001.

3.2.2. Parents.

No significant associations were observed between family meal quantity at baseline and any health or well-being outcomes among parents at follow-up (see Table 4).

Table 4.

Longitudinal Associations Between Family Meal Quantity and Quality and Parent Health and Well-being Outcomes

Family Meal Quantity and Quality at Baseline	Eating and Weight Outcomes at Time 1 Follow-Up, Adjusted for Outcome at Baseline
	Diet Quality	Obesity
	B (95% CI)	PR (95% CI)
Quantity of family meals
Family meal frequency	0.21 (−0.01, 0.44)	1.00 (0.98, 1.02)
Quality of family meals
Interpersonal quality	0.13 (−0.96, 1.22)	1.07 (0.97, 1.18)
Nutritional quality	0.15* (0.01, 0.28)	0.99 (0.98, 1.00)
Family Meal Quantity and Quality! at Baseline	Psychosocial Outcomes at Time 1 Follow-Up, Adjusted for Outcome at Baseline
	Psychological Distress	Severe Anxiety
	B (95% CI)	PR (95% CI)
Quantity of family meals
Family meal frequency	−0.09 (−0.20, 0.02)	0.89 (0.78, 1.02)
Quality of family meals
Interpersonal quality	−0.33 (−0.86, 0.21)	0.84 (0.43, 1.65)
Nutritional quality	0.00 (−0.06, 0.06)	0.98 (0.92, 1.04)

PR, prevalence ratio; CI, confidence interval. Models were adjusted for household race/ethnicity, parent age, parent gender, parent educational attainment, and outcome of interest at baseline. Example interpretation for a B estimate (corresponding to continuous outcome): after adjusting for sociodemographic characteristics and diet quality at baseline, a more positive family meal interpersonal quality predicted a significantly better diet quality score at follow-up by 0.15 points on average. Example interpretation for a PR estimate (corresponding to dichotomous outcome): after adjusting for sociodemographic characteristics and severe anxiety at baseline, each additional family meal per week predicted, on average, 11% lower prevalence of severe anxiety at follow-up, though this association was not statistically significant.

3.2.3. Families.

Family meal quantity at baseline did not demonstrate significant associations with family functioning or chaos at follow-up (see Table 5).

Table 5.

Longitudinal Associations Between Family Meal Quantity and Quality and Family Well-being Outcomes

Family Meal Quantity and Quality at Baseline	Outcomes at Time 1 Follow-Up, Adjusted for Outcome at Baseline
	Family Functioning	Family Chaos
	B (95% CI)	B (95% CI)
Quantity of family meals
Family meal frequency	0.01 (−0.06, 0.08)	−0.03 (−0.07, 0.01)
Quality of family meals
Interpersonal quality	0.14 (−0.18, 0.47)	−0.21* (−0.39, 0.02)
Nutritional quality	−0.02 (−0.06, 0.02)	−0.01 (−0.03, 0.02)

CI, confidence interval. Models were adjusted for household race/ethnicity, parent age, parent gender, parent educational attainment, child age, child gender, and outcome of interest at baseline. Example interpretation for a B estimate: after adjusting for sociodemographic characteristics and family chaos at baseline, a more positive family meal interpersonal quality predicted a significantly lower family chaos score at follow-up by 0.21 points on average.

^*p < .05.

3.3. Associations Between Family Meal Quality and Child, Parent, and Family Health and Well-Being

3.3.1. Children.

Among children, a more positive interpersonal quality of family meals at baseline predicted significantly improved diet quality, reduced emotional problems, and reduced peer relationship problems 18 months later (see Table 3). No significant associations were observed between nutritional quality of family meals and child health or well-being outcomes.

3.3.2. Parents.

Among parents, greater nutritional quality of family meals at baseline predicted significantly improved diet quality 18 months later (see Table 4). No significant associations were observed between family meal interpersonal quality and parent health or well-being outcomes.

3.3.3. Families.

A more positive interpersonal quality during family meals predicted significantly reduced family chaos 18 months later (see Table 5). No significant associations were observed between nutritional quality of family meals and family well-being outcomes.

3.4. Interactions Between Family Meal Quantity and Quality Predicting Child, Adult, and Family Health and Well-Being

One significant interaction between family meal quantity and quality was observed (Figure 1). Specifically, there was a significant interaction between family meal quantity and interpersonal quality of family meals predicting children’s peer relationship problems (p=.02), such that a more positive interpersonal quality during family meals at baseline predicted significantly reduced peer relationship problems at Time 1 follow-up when accompanied by low (defined as the mean minus 1 SD) quantity of family meals (B=−0.40; 95% CI: −0.65, −0.14; p=.003), but not when accompanied by high (defined as the mean plus 1 SD) quantity of family meals (B=0.03; 95% CI: −0.23, 0.29; p=.82).

Figure 1.

Predicted Peer Relationship Problems Among Children at Follow-Up by Baseline Family Meal Quantity and Quality Adjusting for Sociodemographic Characteristics and Peer Relationship Problems at Baseline

Error bars represent standard errors. Low family meal interpersonal quality represents no conversation or with media distractions; high family meal interpersonal quality represents family meals involving conversation without media distractions. Low family meal quantity is the mean −1 standard deviation; high family meal quantity is the mean +1 standard deviation.

4. DISCUSSION

Findings from the current study reinforce and expand prior research on family meal quantity and quality. Current findings support the prior limited research examining family meal quantity and child health and well-being outcomes. Specifically, the protective association found between family meal quantity and reduced prevalence of child obesity provides additional evidence for prior studies in children showing mixed findings between family meal quantity and child weight status.^{20,23, 26,27} These findings suggest that one strategy for decreasing disparities in obesity may be to increase shared meals among families from marginalized ethnic/racial and socio-economic backgrounds. This finding has potential important implications for both researchers and healthcare providers. For example, having family meals may reduce childhood obesity disparities by race/ethnicity through a strengths-based and covert approach that is integrated into family routines—part of families’ everyday lives, rather than overt parenting practices such as weight conversations and controlling food parenting (e.g., restriction) that are often found to backfire when addressing concerns about child weight.^{37,51,66–69}

Findings from the current study also extend previous research on family meals. Regarding family meal quantity, results indicated significant associations between having more family meals and higher diet quality and fewer problematic child eating behaviors, including child food fussiness and food responsiveness. This is a novel finding in the field. Given picky eating is a common complaint among parents ^70–72 and a risk factor for childhood disordered eating behaviors and obesity,^73–76 this finding is potentially highly important for intervention research in that it suggests it may be possible to simultaneously mitigate picky eating and eating in the absence of hunger (i.e., food responsiveness) by increasing the frequency of family meals. Additionally, the finding that family meal quantity was associated with reduced conduct problems in children extends prior research suggesting that family meals are protective for a wide range of child and adolescent health and behavioral indices, from diet quality to emotional and behavioral problems.⁷⁷ The ability to intervene on multiple child behaviors in the same context (i.e., family meals) may be highly appealing to both researchers and clinicians.

Findings from this study regarding family meal quality also advance the knowledge in the field about family meals. The finding that family meal quality, in the form of interpersonal quality at the meal was associated with higher child diet quality and fewer child emotional and peer relationship problems represent new findings that emphasize the importance of communication and few media distractions during the meal to ensure children are nourished physically, emotionally, and socially. This finding expands prior limited studies that have shown better emotional atmosphere during meals and reduced distractions such as watching television were associated with reduced risk for obesity and higher diet quality. ^{9,34,38,78,79} Further underscoring the value of family meal quality, study findings revealed that the interpersonal quality of family meals was associated with reduced family chaos. This finding is also novel and may suggest that routines such as family meals provide a sense of structure, predictability, and safety that reduces the overall level of chaos/disorganization in the home. Prior studies have shown higher family chaos was associated with fewer family meals.⁸⁰

The finding that family meal quality, in the form of nutritional quality during family meals was linked to better parent diet quality 18 months later is new and provides support for the benefits of family meals for all family members, not just children and adolescents. Prior cross sectional research has shown that family meal quantity is associated with better parent diet quality^28,32 and less depressive symptoms, thus this finding additionally suggests the importance of the nutritional quality of the meal in influencing parental diet quality. This finding also emphasizes the potential long-term importance of the healthfulness of the food that is served at family meals, beyond the activity of solely engaging in a family meal.

Our hypothesis that the quantity and quality of family meals would interact synergistically was partially supported. One interaction effect emerged, which suggested that high interpersonal quality at family meals was protective against peer relationship problems for children in families with less frequent family meals. Thus, the interpersonal quality of family meals appears to be especially relevant for social health among children who experience fewer family meals (e.g., one meal per week at baseline), but the predicted differences in peer relationship problems declined as family meal quantity increased (i.e., approached seven family meals per week). A FST lens would suggest that high family meal quality, particularly positive interpersonal interactions, is protective against peer relationship problems because it provides an opportunity to connect and communicate with family members who are emotionally supportive and willing to listen as children talk about their peer relationship problems. It is also possible that high quality family meals may serve as an opportunity for family members to model and teach positive interpersonal skills that may translate into other interpersonal situations (e.g., peer interactions). These opportunities may be especially salient when family meals are infrequent.

Findings from the current study have implications for both clinical and research domains. During well-child visits, clinicians may want to discuss with families whether family meals are occurring and what is taking place during meals given the potential benefits for child, parent, and family health and well-being. In addition, clinicians may want to emphasize the far-reaching protective nature of family meals with multiple child health and emotional well-being outcomes (i.e., weight, eating behaviors, psychosocial outcomes, peer relationships) and support families in identifying steps to increase the frequency and quality of their family meals. Furthermore, given prior research showing that family meals may be more doable than often thought^{2,9,38,81–83} (i.e., can be short, do not need to be from scratch, not necessary to have everyone present) and that even two family meals per week may be protective against future obesity,²² gradual increases in the quantity of family meals and small improvements in the quality of the food served and interpersonal interactions should be encouraged to avoid feelings of being overwhelmed.

Regarding implications for research, it may be worthwhile to target family meal quantity and quality when designing interventions to improve child health and/or well-being as this may yield multiple positive health outcomes and confer benefits beyond the target child. Additionally, findings suggesting that family meals may reduce challenging child eating behaviors (i.e., food responsiveness, food fussiness), emotional problems, peer relationship issues, and conduct problems may be useful when developing family-based interventions aimed at simultaneously modifying multiple health-related behaviors.

Strengths of the study included the longitudinal design, recruitment of racially/ethnically diverse and immigrant/refugee households, use of measures of both family meal quantity and quality, and a broad range of health and well-being outcomes in children, parents, and families. Another strength is the longitudinal nature of the study, which enabled tests of prospective associations.⁸⁴ Study limitations included the use of survey items that may promote socially desirable answers. In addition, the measure of interpersonal quality was limited by a focus on conversation and media and did not include a more comprehensive list of potential interactions that may occur at family meals leading to more or less positive atmosphere. Finally, the population was drawn from one geographic location and generalization of findings may be limited.

5. CONCLUSION

Overall, family meal quantity and quality were found to be independently associated with many child (i.e., reduced obesity prevalence, better diet quality, less food fussiness and food responsiveness, and fewer emotional, peer relationship, and conduct problems) and some parent (i.e., higher diet quality) and family (i.e., less family chaos) health and well-being outcomes. Clinicians may want to emphasize the importance of both family meal quantity and quality during well-child visits, as these longitudinal findings suggest potential benefits for the entire family.

Abbreviations:

FST

Family Systems Theory

ICC

Intraclass Correlation Coefficients

SD

Standard Deviation

Availability of data and materials

Deidentified individual participant data (including data dictionaries) used in the current study will be made available when requested from the first author. Proposals should be submitted to Dr. Jerica M. Berge (jberge@umn.edu).