Data sources, search strategy, and inclusion criteria

No type of meal, publication date, or language restrictions were imposed. Only full-text original articles were included in this review. Quantitative studies of any design were eligible, but qualitative studies were excluded. For the purposes of this review, NS refers to anthropometric measurements and FC refers solely to the intake of food/nutrients and the quality of the diet; thus, studies that associated other aspects of eating behaviors with FFM were not considered. Studies were considered eligible regardless of whether data on FFM was self-reported by the adolescents or reported by parents/legal guardians. Studies were not required to cite the association between these variables as a main objective to be included in the review. Studies involving older children or young adults in addition to adolescents were also considered eligible for inclusion. Investigations of clinical populations or individuals with eating disorders, studies where exposure or outcomes did not occur during adolescence, and studies in which the outcome was not the frequency but rather the quality of family meals were excluded from this review.

Studies were identified by searching 5 electronic databases (PubMed, Web of Science, Scopus, BVS Brazil, and Adolec) and gray literature (Google Scholar) and by scanning reference lists of included studies. Google Scholar searches were by relevance and limited to the first 200 citations. Three sets of MeSH terms were used to account for FFM, NS, and FC and needed to be displayed in the title or abstract. The search strategy was initially developed for PubMed (Fig 1) and later adapted for use in all other databases. No filters were applied to any database in the initial search, which was run in August 2019. An updated literature search was performed in May 2020 in all databases using the same search strategy, but with a filter for year of publication (2019–2020).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PubMed search strategy.

https://doi.org/10.1371/journal.pone.0239274.g001

Study selection process

Two independent reviewers (R1 and R2) screened titles and abstracts identified by the initial search. Disagreements between the 2 reviewers were resolved by consulting a third reviewer (R3) for arbitration. Duplicates were removed, and the remaining studies were organized using Mendeley Ltd. Full text was obtained for all studies meeting the inclusion criteria and were read independently by 2 reviewers (R1 and R2). Disagreements were resolved by discussion and consensus among R1, R2, and R3, or by consulting an expert (R5). The reference lists of all studies selected for full-text review were hand-searched by R3, and decisions to include additional studies were made in a meeting involving all the aforementioned reviewers. In the update, the titles and abstracts of the articles retrieved were screened by 3 reviewers (R1, R2 and R3) simultaneously, and disagreements were resolved in a discussion meeting. Subsequently, all studies meeting the inclusion criteria were independently read in full by the 3 reviewers (R1, R2 and R3). There were no disagreements at this stage.

Extraction, risk of bias, and data analysis

The following data were extracted from each study and included in a table designed specifically for this review by 2 of the reviewers (R2 and R3): study characteristics (year of publication, funding sources, and country of origin); sample characteristics (age, recruitment process, and ethnicity); family meal characteristics (type of meal evaluated and FFM); and outcome characteristics (main results, variables measured, and type of assessment). A third reviewer (R1) checked the data for consistency and accuracy.

Risk of bias was assessed by one reviewer (R1), while a second reviewer (R2) reviewed the ratings for consistency and accuracy. Studies were evaluated using the Newcastle-Ottawa scale (NOS) for cohort and cross-sectional studies. The NOS is increasingly used as a measure to assess the quality of nonrandomized trials and provides a separate version for each of these study designs [15, 16]. Published thresholds were used to convert NOS scores into Agency for Healthcare Research and Quality (AHRQ) standards (good, fair, and poor). Because item 4 in the selection domain of the NOS for cohort studies was rated as ‘not applicable’ for all included articles, those with at least 4 stars in other domains were considered of good quality. Investigations with both longitudinal and cross-sectional components were analyzed twice, once for each study design.

Data were compiled into 2 summary tables (NS and FC). A meta-analysis was not possible due to the methodological heterogeneity caused by variations in exposure and outcome measurements across studies, in addition to the absence of confidence intervals and odds ratios in studies that did not investigate the association between FFM and FC or NS as the main objective. A narrative synthesis of extracted findings was conducted by grouping secondary variables into the following 5 categories: dietary habits, lifestyle and social factors, psychological factors, family meal environment, and adolescent health. These findings were discussed by focusing on the association of these variables with FFM, rather than on FFM itself.

Search results, design & quality assessment

The initial search yielded 2001 results, of which 47 were included. The search update yielded 329 results, of which 3 were eligible for inclusion, for a total of 50 articles included in this review. The study selection process, including the initial search and the update, is shown in Fig 2. Twenty-five studies investigated the association between FFM and NS [12, 17–40] and are summarized in Table 1. The association between FFM and FC was investigated in 32 studies, which are summarized in Table 2 [11, 17, 22, 23, 27, 30, 38, 39, 41–64]. Seven studies [17, 22, 23, 27, 30, 38, 39] focused on both NS and FC and were included in both tables. Thirty-four were cross-sectional studies [11, 17, 18, 21, 22, 25, 27, 29, 31–35, 37–41, 43–46, 48, 50, 51, 53, 55–62], 12 were longitudinal [19, 20, 24, 26, 28, 29, 31, 42, 47, 49, 63, 64], and 4 studies analyzed both cross-sectional and longitudinal data [12, 36, 52, 54].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Study selection process.

https://doi.org/10.1371/journal.pone.0239274.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Description of studies investigating the association between family meal frequency and adolescent nutritional status.

https://doi.org/10.1371/journal.pone.0239274.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Description of studies investigating the association between family meal frequency and adolescent food consumption.

https://doi.org/10.1371/journal.pone.0239274.t002

The cross-sectional and longitudinal components of studies with both types of data were evaluated separately for risk of bias. According to the NOS, 35 studies were rated as having good quality (27 cross-sectional and 8 longitudinal) [12, 17–26, 32–38, 42–45, 47, 48, 50–52, 54, 56, 58–61, 63, 64], whereas 19 were of fair quality (15 cross-sectional and 4 longitudinal) [11, 20, 27–31, 36, 39–41, 46, 49, 52–55, 57, 62]. In most cases, cross-sectional studies were rated as having fair quality because of lack of justification of sample size, no calculation of the response rate, or no description of characteristics of responders and non-responders, and longitudinal studies because of written self-report questionnaires to investigate exposure, follow-up rate <80%, or no description of participants lost to follow-up.

Overview of included studies

The studies included in this review were published between 2000 and 2020, with 23 published in the past 5 years [17, 22, 25–29, 33, 37, 41, 43, 44, 48, 49, 51, 53, 55–61]; of these, 8 were published between 2019 and 2020 [17, 20, 22, 26, 41, 43, 51, 58]. Thirty-one studies were conducted in North America [11, 12, 19, 20, 23, 24, 27–32, 34, 35, 36, 41–47, 49, 52, 53, 55, 56, 61–64], and 27 of these were conducted in the United States. An additional 6 studies were performed in Europe [21, 22, 38, 41, 50, 60], 4 in Oceania [18, 39, 40, 54], and 4 in Latin America [26, 48, 58, 59]. Three studies were performed in Asia [25, 51, 57], 1 in Africa [33], and 1 in the Middle East [17].

Main outcomes of FFM & NS

Ten studies investigated the association between FFM and BMI classifications. Of these, 6 identified a positive relationship between FFM and better NS. Alamri (2020) examined the influence of family meal type (breakfast, lunch, and dinner) on food intake and BMI of Saudi adolescent girls and found that eating a family lunch or dinner was negatively associated with adolescent BMI; however, no significant association was found between family breakfast and adolescent BMI [17]. Chang & Halgunseth (2014) examined the influence of acculturation, ethnicity, interactions among family meals, parental control, and FFM on changes in NS in Hispanic and Caucasian adolescents in the United States over approximately 4 years. The results showed that non-Hispanic white adolescents were more likely to maintain a healthy weight, not to purchase school lunch, and to eat more family meals. There was also an association between high FFM and a negative change in NS (normal to overweight) in Hispanic adolescents with lower parental behavioral control (high behavioral control and use of power assertive parenting techniques to address child misbehavior are part of the demandingness dimension of parenting style framework) [19]. Farajian et al. (2014) identified FFM as 1 of the 5 main predictors of overweight in childhood, as having frequent family meals was associated with a reduced risk of overweight and obesity [21]. Larson et al. (2013) found that a higher frequency of family breakfasts was associated with a lower risk of overweight [30]. Kubik et al. (2009) showed that obese students were more likely than their normal-weight peers to report no family meals in the previous week [29]. Sen (2006) assessed FFM and NS in individuals of different ethnicities and found that, for white adolescents, high FFM was associated with reduced weight in 1997 and with a reduced likelihood of becoming overweight and increased odds of ceasing to be overweight in 2000 [34]. Vik et al. (2016) evaluated the association between eating the 3 main meals with family members and NS in adolescents and showed that having regular family breakfast, but not lunch or dinner, was inversely associated with overweight [37]. The other 3 studies investigating these variables did not identify any significant association between them [20, 32, 40].

Eight studies evaluated the association between FFM and NS as defined by BMI. Six of these studies identified a negative association between exposure and outcome. Fulkerson et al. (2008) evaluated NS and FFM in adolescents over the course of 5 years and reported that, although these variables were not longitudinally associated, significant inverse relationships occurred between FFM and overweight in female adolescents in all cross-sectional models [12]. Goldfield et al. (2019) found that higher FFM was associated with lower BMI only in female participants [24]. Haghighatdoost et al. (2017) observed that family dinner frequency was inversely related to obesity in adolescents [25]. Hassan et al. (2019), in a 3-year longitudinal study, showed that male participants who did not have breakfast with their families had the largest increases in BMI over the course of the study, while the largest reductions in percent body fat were observed in male adolescents who reported an intermediate frequency of family breakfasts [26]. Taveras et al. (2005), in cross-sectional analyses, identified a lower likelihood of becoming overweight in participants who ate family dinner on “most days” or “every day” than in those who ate family dinner “some days or never”; no significant associations were observed in longitudinal analyses [36]. In the studies conducted by Babajafari et al. (2011) [18], Utter et al. (2008) [38], and Utter et al. (2013) [39], no significant associations were observed between FFM and NS in adolescents.

Seven studies also evaluated the association between FFM and NS using BMI z-scores or age-specific percentiles, and all of them identified positive associations between FFM and lower odds or prevalence of overweight. Frank et al. (2019), aiming to show current family meal patterns of children and adolescents aged 6 to 17 years living in Germany and to investigate associations with sociodemographic characteristics, BMI, and dietary behavior, found that overweight participants aged 12–17 years ate breakfast, afternoon snacks, or dinner together with their families less frequently than their non-overweight peers [22]. Fulkerson et al. (2009) reported that adolescents who did not eat dinner with their families in the week before the study were almost 3 times more likely to be overweight [23]. Horning et al. (2016), evaluating different measures of FFM, as reported by parents and adolescents, and the relationship between these variables and BMI z-scores, found, in adjusted analyses, that measures of sitting and eating together were inversely related to BMI z-scores in adolescents [27] In another study, Horning et al. (2017) evaluated the moderating role of FFM in the relationship between meal context and NS of parents and adolescents, and the results suggested that FFM amplifies the relationship between healthy eating behavior and BMI in the context of family meals [28]. Larson et al. (2013) showed that infrequent family meals were consistently associated with higher BMI z-scores in both boys and girls [31]. Sedibe et al. (2018), in cross-sectional analyses, showed that eating the main meal with the family on “some days” or “nearly every day” was 1 of the 2 main variables responsible for an increased risk of overweight and obesity only in black or rural preadolescents [33]. The only study to use BMI percentiles was conducted by Smith Price et al. (2009), who found that more frequent family meals led to improvements in NS over time [35].

Main outcomes of FFM & FC

Most studies (n = 24) used food-frequency questionnaires that had been previously validated or developed specifically for the study [11, 22, 23, 30, 38, 39, 42–52, 54, 55, 57–61]. The remaining studies used 24-h dietary records [17, 27, 41, 53, 56, 62–64].

Using a food-frequency questionnaire, Fulkerson et al. (2009) analyzed breakfast content, dietary habits, fruit and vegetable consumption, and unhealthy food intake and found that family dinner frequency was positively associated with breakfast frequency and daily fruit intake, but not with fast-food restaurant use, regular soft drink consumption, vegetable consumption, combined fruit and vegetable consumption, or high-fat food intake [23]. Frank et al. (2019), in a subsample from the German Health Interview and Examination Survey for Children and Adolescents (KiGGS/2014-2017), investigated an association between FFM and food intake or the frequency of consumption of food groups and found that the daily consumption of sugary drinks is lower in adolescents aged 12–17 years who have breakfast with their families every day/often than in those who rarely/never have breakfast with the family. Analyses for other food groups, including fruit, vegetables, water, confectionery/salty snacks, milk products, meat/sausages, cereal products, fast food, and fish, showed no significant associations [22]. Larson et al. (2013) evaluated the association between frequency of family breakfasts and food intake and found that adolescents who ate more meals with their families were more likely to have positive dietary habits, including increased consumption of fruit, whole grains, fiber, and potassium, as well as reduced consumption of sugar-sweetened beverages [30]. In the studies conducted by Utter et al. (2008) and Utter et al. (2013), a positive association was identified between higher fruit and vegetable intake and frequency of family breakfast [38, 39].

Burgess-Champoux et al. (2009) examined the longitudinal association between participation in family meals and dietary habits in adolescence over a 5-year period. Male and female adolescents who ate regular family meals (at least 5 per week) had higher intakes of vegetables, calcium-rich foods, fiber, and several vitamins and minerals [42]. Conlon et al. (2019) assessed the influence of family environment on obesity-related health behaviors in adolescents and showed that participants with higher FFM were more likely to eat fruits, but they did not differ from their peers in the consumption of vegetables or sugar-sweetened beverages. Additionally, adolescents whose parents reported a higher frequency of television watching during family meals were likely to eat less fruit [43]. Demissie et al. (2015) classified participants’ dietary habits as healthy or unhealthy based on the intake of different food groups and found that healthier eating habits were associated with higher FFM [44].

Feldman et al. (2007) reported that adolescents who watched television during family meals were more likely to have a poorer diet than those who did not. Television watching during family meals was associated with lower intakes of vegetables, grains, and dairy products and with higher intakes of soft drinks and fried foods. Additionally, the results showed that adolescents eating regular family meals while watching television had better dietary intake than those not eating regular family meals [45]. Fink et al. (2014) also found that having 5 or more family meals per week was associated with higher intakes of vegetables and fruit in participants aged 12–17 years [46]. Gillman et al. (2000) reported that a higher frequency of family dinners was associated with healthier eating habits, including higher intakes of fruit, vegetables, fiber, and micronutrients and lower intakes of fried foods, soft drinks, and saturated and trans fats. In addition, FFM was associated with lower glucose levels, although it was not significantly associated with the intake of red meat or snack foods [47]. Hong et al. (2019), using data from a national Korean survey to compare adolescents who had breakfast with their families with those who had this meal alone, performed analyses for the association between FFM and nutritional quality, food intake, meal frequency, and energy intake and found that the food intake of carbohydrates and iron, the average Dietary Diversity Score, and the number of servings of grains and vegetables were significantly higher in the family breakfast group. However, the consumption of milk and dairy products was higher in the group of participants eating breakfast alone [51].

Giray & Ferguson (2018), examining the association between the frequency and quality of family meals and nutritional outcomes in young Jamaicans, found no significant associations between these variables [48]. Granner & Evans (2011), evaluating individual characteristics, social factors, family environment and their relationship to fruit and vegetable intake in adolescents, showed, in multivariate analyses, a positive association between the frequency of family dinners and daily fruit and vegetable intake, as evidenced by differences between adolescents who ate fewer than 3 portions and those who ate 3–4 portions of fruit and vegetables per day, and between those who ate fewer than 3 portions and those who ate more than 5 portions per day [49]. Haapalahti el at. (2003) found a significant association between FFM and healthier eating habits, such as lower intakes of fast foods and sweets by adolescents [50]. Larson et al. (2007), investigating FFM in adolescence and its relationship to food intake in young adulthood, showed a significant association of FFM with a higher quality diet, higher intakes of fruit, vegetables, and micronutrients, and lower intake of soft drinks [52]. Alamri (2020), examining the influence of the type of family meal (breakfast, lunch, and dinner) on the food intake of female adolescents, found that family breakfast was positively associated with the consumption of dairy products and proteins, family lunch was positively associated with eating vegetables and proteins, and family dinner was positively associated with the intake of fruit, vegetables, dairy products, and whole grains [17].

Leech et al. (2014) reported that daily family dinners were associated with healthy food habits, including higher consumption of fresh and dried fruits, vegetables, low-fat milk, and water and lower intake of energy-dense foods (sweet and savory foods, high-energy beverages), but only in male adolescents [54]. Lipsky et al. (2015) examined tendencies and changes in eating behaviors during the transition between adolescence and early adulthood and found that the intake of fruits, vegetables, and whole grains in this age range was positively associated with FFM [55]. Makansi et al. (2018) also evaluated eating habits and factors associated with fruit and vegetable intake, but no significant relationships were identified between FFM and fruit or vegetable consumption [57]. Martins et al. (2019) showed that eating at least 5 family meals per week was positively associated with the intake of beans, fruits, and vegetables (healthy diet score) and negatively associated with the consumption of sweets, ultra-processed salty foods, and fried salty snacks (unhealthy diet score) [58].

Neumark-Sztainer et al. (2003) assessed the patterns of family meals and their relationship to food and nutrient intake in a population of adolescents in the United States and found that adolescents with more frequent family meals had higher intakes of fruits, vegetables, grains, and calcium-rich foods. FFM was also negatively associated with the intake of soft drinks and snack foods [11]. Oliveira et al. (2018) reported that participants who ate more meals at home with their parents had a healthier diet and higher intakes of beans, vegetables, and milk, in addition to consuming smaller amounts of fried foods, sweets, and soft drinks on weekdays [59]. Totland et al. (2017) reported that lower frequency of family dinners was associated with lower vegetable intake among adolescents [60]. Walton et al. (2018) found that frequent family dinners were associated with healthier dietary habits, including higher fruit and vegetable intake and lower consumption of fast food and takeout in both boys and girls, as well as lower intake of sugar-sweetened beverages in boys [61].

Horning et al. (2016) identified a significant association of the frequency of family dinners with dietary quality and fruit and vegetable intake, as assessed by the Healthy Eating Index (HEI) [27]. Dietary quality and the intake of fruits and vegetables at dinner were also assessed by Arcan et al. (2019), who found that individuals with higher FFM were more likely to fill half their plate with fruits and vegetables, although they did not differ from their peers in terms of dietary quality [41]. Larson et al. (2016), investigating family meal experiences and fast-food purchasing among adolescents, evaluated several indicators of diet quality, but the only significant associations were found between the frequency of family breakfast and fruit intake in males, and refined grain intake in females [53]. Lipsky et al. (2017) showed that higher FFM was associated with higher scores on 3 diet quality indicators (HEI, Whole Plant Foods Density, and Empty Calories) [56]. Woodruff et al. (2009) showed that a higher frequency of family dinners was significantly associated with lower soft drink consumption and breakfast on the day of the study among adolescents [62]. Woodruff et al. (2010), investigating the association between FFM and overall diet quality, reported that adolescents who had 6 to 7 family meals per week had higher HEI scores than their peers [63]. Lastly, in another study, Woodruff et al. (2014) examined the association between family dinner frequency and number of meals and snacks per day and found that adolescents who had regular dinner with their families were likely to eat more meals and snacks on a daily basis as well as to consume fewer calories per day than those eating dinner with friends or alone [64].