SEARCH

SEARCH BY CITATION

Keywords:

  • family eating;
  • maternal attitudes;
  • adolescence overweight

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Objectives: To determine whether maternal attitude towards the family regularly eating together and maternal report of how often the family eat together are associated with adolescent offspring overweight.

Research Methods and Procedures: A cross-sectional mother-child-linked analysis was carried out using 14-year follow-up data from a population-based prospective birth cohort of 3795 children (52% males) who were participants in the Mater-University study of pregnancy, Brisbane, Australia. Maternal reports on family eating pattern reported at age 14 were used.

Results: The prevalence of overweight at age 14 was 24.1% (95% confidence interval (CI), 22.3, 26.1) for males and 27.1% (CI, 25.1, 29.2) for females. The majority of mothers (78%) reported that the family ate together at least once a day, but only 43% reported that they felt that family eating together was important. The offspring of women who felt that the family eating together was not important had increased odds of being overweight at age 14 (odds ratio, 1.27; 95% CI, 1.05, 1.53) in age- and sex-adjusted models. Adjustment for potential confounding factors had no substantive effect on the association. There was no association between maternal reports of how often the family actually did eat together and overweight at age 14 in the offspring.

Discussion: These findings suggest that maternal attitude towards family eating patterns, but not maternal report of how often the family do eat together, are associated with childhood overweight status. Maternal attitude towards family eating (as opposed to a report of actual frequency at one time-point) may reflect broader maternal influences (beyond family eating pattern) on their child's diet and eating patterns over a long time course.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

The population prevalence of childhood obesity has increased by about 3-fold in most industrialized countries over the last 10 to 20 years (1). These trends are likely to have major public health consequences (1). Obesity tracks from childhood to adulthood, and there is emerging evidence that the precursors for diabetes and cardiovascular disease, some of which will be irreversible, already exist in obese children (1, 2, 3, 4, 5, 6). However, it is unclear how to prevent further increases in the population prevalence of childhood overweight (7).

The idea that one needs to consider the environment in which a child or adolescent eats when thinking of strategies to prevent obesity is not new, and the notion that family eating habits contribute to obesity problems in childhood is a common perception in the media. However, surprisingly little epidemiological evidence exists to support or refute these ideas. Several studies have found that children and adolescents who eat dinner with family members more often are more likely to eat fruits and vegetables and are less likely to eat high-fat foods, convenience foods, and sweets and to drink large amounts of carbonated drinks (8, 9, 10, 11, 12, 13). However, whether these dietary effects lead to appropriate weight gain and size during childhood and into adulthood is unclear. Strong graded associations have also been found between hours of television watching and the risk of being overweight among children (14, 15). This may be, in part, related to reduced levels of activity among those who watch more television but may also be due to the fact that children watching more television are less likely to participate in family meals and are more likely to snack and eat junk food while watching the television (15, 16, 17).

The aim of this study was to assess the effect of maternal attitude towards whether the family ate together and maternal report of how often they ate together on the odds of being overweight in a cohort of Australian adolescents who have been followed-up since their birth in a public hospital in Brisbane in the 1980s.

Research Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Participants

The Mater-University study of pregnancy and its outcomes (MUSP)1 is a longitudinal study of women and their offspring, who received antenatal care at a major public hospital (Mater Misericordiae Hospital) in South Brisbane between 1981 and 1984 (18). In total, 7223 mothers agreed to participate, delivered a live singleton baby who was not adopted before leaving the hospital, and completed both initial phases of data collection; these mothers and their offspring form the MUSP prospective cohort.

The mothers and children have been followed-up prospectively, with maternal questionnaires being administered when their children were 6 months, 5 years, and 14 years. In addition, at 5 and 14 years, detailed physical, cognitive, and developmental examinations of the children were undertaken, and at 14 years, the children completed health, welfare, and lifestyle questionnaires. Of the original 7223 participants, BMI (weight in kilograms divided by height in meters square) of 3795 (53%) was measured at 14 years. Children who could not participate were more likely to be from families with low income at birth, to have mothers who smoked throughout their pregnancy, and to have mothers and fathers with lower educational attainment (19).

Measurements of Exposures and Outcomes

In this study, the main outcome in all analyses was the adolescent's BMI at the 14-year follow-up. In all assessments, the average of two measures of the participant's weight while lightly clothed, with a scale accurate to 0.2 kg, was used. A portable stadiometer was used to measure the height. Overweight was defined according to standard definitions derived from international surveys by Cole et al. (20). Thus, using the standard definition, overweight at age 14 was defined as a BMI exceeding 22.62 kg/m2 for a male participant and 23.34 kg/m2 for a female participant. These values are equivalent to exceeding 25 kg/m2 in adulthood (20). These BMI cut-off points are equivalent to the 85th percentiles on the CDC 2000 scale (21). Too few of the participants fell into the obese categories for meaningful analyses; therefore, we classified children as either overweight or obese (equal to or above the overweight threshold) or normal weight (below the overweight threshold). Those who are either overweight or obese are referred to as overweight throughout the remainder of the paper.

Two main exposures considered were mother's report on how often the family ate together and her attitude towards the family eating together. At the 14-year follow-up, the mothers were asked how often her family ate together, with the given response options: at least once a day, a few times a week, and about once or less than once a week. Each mother was also asked how important she felt it was that the family ate together and was given the response options: very important, quite important, and not really important. This was a cross-sectional investigation based on the 14-year follow-up data of a prospective birth cohort. Because information on family eating together or maternal attitude towards this were not collected at the 5-year (or earlier) follow-up, it was not possible to conduct a prospective analysis.

In some analyses (due to small numbers), maternal report of how often the family ate together was collapsed into two categories (at least once a day vs. few times/once/less than once a week). Similarly, in some analyses, maternal attitude towards the family eating together was collapsed into very important vs. quite important or not really important.

The following potential confounders were taken into consideration: child's sex and age, maternal parity, gross family income at 14-year follow-up (low, <Australia $15, 599; medium, Australia $15, 600 to $31, 148; high, Australia, $31, 149 or more), maternal education at birth (did not complete secondary school, completed secondary school, completed further/higher education), and race (white, Asian, and Aboriginal-Islander). All of these variables were obtained from the obstetric records or from interviews with the mothers during pregnancy or around the time of the birth of the child. In addition, the following, which were obtained from questionnaires completed by the mother at the 14-year follow-up (18), were also considered as potential confounders: the importance of whether food that the mother chose for her children was fattening (very important, quite important, and not important); her response to food frequency questions relating to the child's diet regarding fast food, soft drinks, and red meat (all with response options of rarely or never, at least two or three times a week, most days); her report of the amount of time the child spent watching television (<1 hour per day, 1 to <3 hours per day, 3 to <5 hours per day, and 5 or more hours per day); and time spent on sports or exercise per week (4 to 7 days per week, 0–3 days per week).

To assess whether the child's overweight status had affected maternal attitude towards family eating, we also adjusted for whether the child was overweight at age 5. Here, we were addressing the possibility that if a child had been overweight at an earlier stage, this may have affected family eating patterns and maternal attitude towards them; being overweight at age 5 is also likely to be strongly associated with being overweight at age 14. At age 5, being overweight was defined, according to Cole et al.'s (20) international surveys, as having a BMI > 17.42 kg/m2 for male participants and 17.15 kg/m2 for female participants.

Statistical Analyses

Logistic regression was used to estimate the age-adjusted prevalence of overweight at age 14 and association of exposures and other covariates with the child being overweight at age 14. It was also used to estimate the age-adjusted prevalence of maternal attitude towards eating together and mothers’ report on how often the family ate together and associations of covariates with these exposures at age 14.

In the final analyses, a series of multiple logistic regression was used to further assess the association of maternal attitude towards family eating together and their offspring being overweight at age 14 and the association of maternal report on how often family had meal together and their offspring being overweight at age 14. In the first model, age (continuous variable) and sex were included as confounders. In the second model, being overweight (binary variable) at age 5 was added. In the third model parity (three-level categorical variable), gross family income in the year of birth (three-level categorical variable), maternal education (three-level categorical variable), and maternal race (three-level categorical variable) were added, and in the final model, maternal attitude towards the child's consumption of fatty foods and her report of the child's consumption of fast food, soft drinks, and red meat, amount of television watching, and participation in sports and exercise (all are categorical variables) were added.

To assess whether non-response biased our results, we used logistic regression (response vs. non-response as outcome) to determine weights for each individual using the inverse probability of response (22). Response and non-response categories were defined by our inclusion criteria; i.e., a responder (someone included in the analyses) was a child whose mother had responded to family eating pattern at age 14 and on whom there were height and weight measurements (for the calculation of BMI) at age 4 years. A non-responder was any other child in the original birth cohort. Non-response was predicted mainly by mother's age at birth, family income, maternal education, and ethnic groups of the parents. The individual weighting factor for these covariates (their inverse probability) was used as a sample weighting adjustment into the multivariable logistic regressions. For instance, using the predictive model, if we found that the probability of non-response was 0.65 for a mother who did not complete secondary school, the inverse probability of weight for her was 1.54. All analyses were undertaken using Stata version 8.0 (Stata Corporation, College Station, TX).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

The age-adjusted prevalence of being overweight at age 14 was 24.1% [95% confidence interval (CI), 22.3, 26.1] for males and 27.1% (25.1, 29.2) for females. Although the majority of the mothers (78%) reported that they had a family meal at least once a day, just 43% of mothers stated that they thought that eating together as a family was very or quite important. Figure 1 shows the comparison of overweight prevalence between different combinations of the categories of the two exposures. For all three categories of maternal report of the frequency of eating together, children whose mothers thought that eating together was important had a lower prevalence of being overweight or obese than those whose mothers did not think it was important. However, maternal report of frequency of eating together did not seem to importantly influence the prevalence of overweight.

image

Figure 1. Comparison of overweight prevalence (percentage) between different combined categories of maternal attitude towards the family eating together (very important, quite important, and not really important) and maternal report of how often the family ate together (at least once a day, a few times a week, and once or less than once a week) at age 14. Error bars represent the 95% CIs for the prevalence of overweight.

Download figure to PowerPoint

Table 1 shows the age-adjusted associations of characteristics with being overweight at age 14. There was no association between maternal report of how often the family ate together and whether the child was overweight. Children of mothers who felt that eating together was not important were more likely to be overweight at age 14. Females were more likely to be overweight than males and those from lower income families, whose mothers had lower levels of education, whose mothers thought that eating fattening foods was not important, and whose mothers reported that their offspring consumed a lot of soft drinks and red meat were more likely to be overweight. Maternal report of high frequency of television watching and low rates of participation in sports and exercise was associated with increased odds of being overweight, whereas maternal report of consumption of fast foods was not associated with being overweight. There was a tendency for children of Asian mothers to be less likely to be overweight than those of white or Aboriginal-Islander mothers. By far, the strongest predictor of being overweight at age 14 was being overweight at age 5 [odds ratio (OR), 8.07; 95% CI, 6.55, 9.95].

Table 1. . Age-adjusted associations of maternal attitude towards family eating, maternal report of how often the family ate together, and other characteristics with being overweight at age 14
  Overweight at age 14 
 N% (95% CI)pAge-adjusted ORs (95% CI) of being overweight or obese at age 14
How often family ate meal together
 At least once a day291925.8 (24.2, 27.4) 1.00
 Few times/once/less than once a week83824.7 (21.9, 27.7)0.5430.95 (0.79, 1.13)
Maternal attitude towards family eating together
 Very important163123.2 (21.2, 25.3) 1.00
 Not important212227.3 (25.4, 29.2)0.0031.25 (1.07, 1.4)
Sex
 Male197824.1 (22.3, 26.1) 1.00
 Female181727.1 (25.1, 29.2)0.041.17 (1.01, 1.35)
Overweight status at age 5
 Normal244317.4 (15.9, 18.9) 1.00
 Overweight50663.0 (58.8, 67.1)<0.0018.07 (6.55, 9.95)
Family income
 Low30029.0 (24.1, 34.3) 1.00
 Medium138327.5 (25.2, 29.9) 0.92 (0.70, 1.21)
 High199823.3 (21.5, 25.3)0.0070.74 (0.56, 0.97)
Mother's education
 Did not complete secondary61428.6 (25.2, 32.3) 1.00
 Completed secondary242925.0 (23.3, 26.8) 0.83 (0.68, 1.01)
 Completed higher72824.8 (21.8, 28.1)0.150.81 (0.64, 1.03)
Race
 White336625.5 (24.1, 27.0) 1.00
 Asian14718.3 (12.9, 25.4) 0.65 (0.43, 0.99)
 Aboriginal-Islander17128.0 (21.8, 35.2)0.081.14 (0.81, 1.61)
Parity
 1152226.2 (24.0, 28.4) 1.00
 2116423.3 (21.0, 25.8) 0.85 (0.71, 1.02)
 3+110426.9 (24.3, 29.6)0.101.03 (0.87, 1.23)
Attitude towards fattening food
 Very important37217.4 (13.9, 21.6) 1.00
 Quite important197526.1 (24.2, 28.1) 1.67 (1.25, 2.22)
 Not important137627.1 (24.8, 29.5)<0.0011.76 (1.31, 2.36)
Fast food
 Never or rarely276025.4 (23.8, 27.1) 1.00
 Most days or 2/3 days per week94626.4 (23.7, 29.3)0.541.05 (0.89, 1.25)
Soft drinks
 Never or rarely129422.1 (19.9, 24.4) 1.00
 2/3 days per week126324.6 (22.3, 27.0) 1.16 (0.97, 1.39)
 Most days117030.7 (28.2, 33.4)<0.0011.60 (1.33, 1.91)
Red meat
 Never or rarely24519.6 (15.1, 25.0) 1.00
 2/3 days per week159124.5 (22.5, 26.7) 1.53 (1.10, 2.13)
 Most days188627.2 (25.3, 29.3)0.021.33 (0.95, 1.86)
Sports
 0 to 3 days per week181830.0 (28.0, 32.3) 1.00
 4 to 7 days per week194021.2 (19.5, 23.1)<0.0010.62 (0.54, 0.72)
Watching television
 <1 hour per day30521.9 (17.6, 26.9) 1.00
 1 to <3 hours per day98023.8 (21.3, 26.6) 1.12 (0.82, 1.52)
 3 to <5 hours per day103728.4 (25.7, 31.2) 1.42 (1.05, 1.92)
 5 or more hours per day141425.4 (23.2, 27.7)0.041.22 (0.90, 1.64)

Table 2 shows the age-adjusted associations of covariates with a maternal attitude that family eating together was not important and her report of frequency of eating together as a family. Mothers with lower levels of education were more likely to say that family eating together was not important. Asian mothers and Aboriginal-Islander mothers were less likely than white mothers to feel that eating together was not important. Mothers who reported feeling that family eating together was not important were also more likely to report that their child consumed a lot of fast food and soft drinks and watched more television. Maternal report of frequency of eating together was associated only with sex, parity, fast food, and red meat.

Table 2A. . Age-adjusted associations of maternal, family, and child characteristics with maternal attitude towards family eating and maternal report of how often the family ate together
  Mother feels family eating together is not important Mother reports family has meal together few times/once/less than once a week 
 N% (95% CI)pAge-adjusted ORs (CI) of mother feeling family eating together is not important% (95% CI)pAge-adjusted ORs (CI) of mother reporting family has had meal together few times/once/less than once a week
Sex
 Male197855.5 (53.2, 57.6) 1.0020.8 (19.3, 22.6) 1.00
 Female181757.7 (55.4, 60.0)0.1601.10 (0.96, 1.25)24.0 (22.0, 26.0)0.0181.20 (1.03, 1.40)
BMI at age 14
 Normal282455.2 (53.3, 57.03) 1.0022.5 (20.9, 24.2) 1.00
 Overweight97160.05 (57.4, 63.5)0.0041.25 (1.07, 1.45)21.8 (18.4, 25.6)0.5440.95 (0.79, 1.13)
Family income
 Low30058.6 (52.9, 64.1) 1.0021.8 (17.5, 26.9) 1.00
 Medium138357.9 (55.3, 60.1) 0.98 (0.76, 1.26)20.6 (18.6, 22.8) 0.94 (0.69, 1.27)
 High199855.9 (53.7, 58.1)0.4620.90 (0.70, 1.16)23.7 (21.9, 25.7)0.0911.12 (0.84, 1.51)
Mother's education
 Did not complete secondary61460.2 (56.1, 64.0) 1.0025.5 (22.1, 29.1) 1.00
 Completed secondary242957.6 (55.6, 59.6) 0.90 (0.75, 1.08)22.3 (20.7, 24.0) 0.84 (0.68, 1.03)
 Completed higher72849.8 (46.1, 53.4)<0.0010.66 (0.53, 0.82)20.0 (17.2, 23.0)0.0580.73 (0.57, 0.95)
Race
 White336657.5 (55.8, 59.1) 1.0022.2 (20.8, 23.6) 1.00
 Asian14743.1 (35.2, 51.3) 0.56 (0.40, 0.79)22.1 (16.1, 29.5) 1.00 (0.67, 1.49)
 Aboriginal-Islander17152.7 (45.1, 60.3)0.0020.82 (0.60, 1.13)26.4 (20.2, 33.7)0.4751.07 (0.88, 1.29)
Parity
 1152255.4 (52.9, 57.9) 1.0018.4 (16.5, 20.4) 1.00
 2116456.7 (53.9, 59.6) 1.06 (0.91, 1.24)23.3 (21.0, 25.9) 1.35 (1.12, 1.64)
 3+110457.9 (54.9, 60.8)0.451.10 (0.94, 1.29)26.5 (24.0, 29.2)<0.0011.60 (1.33, 1.93)
Attitude towards fattening food
 Not important37265.2 (60.4, 70.1) 1.0022.7 (18.7, 27.2) 1.00
 Quite important197561.1 (58.9, 63.2) 0.83 (0.66, 1.05)23.2 (21.4, 25.1) 1.03 (0.79, 1.35)
 Very important137647.8 (45.1, 50.4)<0.0010.48 (0.38, 0.61)20.8 (18.8, 23.1)0.2490.90 (0.68, 0.18)
Fast food
 Never or rarely276054.0 (52.1, 55.8) 1.0020.8 (19.3, 22.3) 1.00
 Most days or 2/3 days per week94664.1 (61.0, 67.1)<0.0011.52 (1.31, 1.77)26.9 (24.2, 29.8)<0.0011.40 (1.18, 1.66)
Soft drinks
 Never or rarely129455.2 (52.5, 57.9) 1.0023.5 (21.3, 25.9) 1.00
 2/3 days per week126354.1 (51.3, 56.8) 0.95 (0.81, 1.11)22.5 (20.2, 24.9) 0.94 (0.78, 1.13)
 Most days117060.8 (58.0, 63.6)<0.0021.25 (1.06, 1.47)20.9 (18.7, 23.3)0.2550.85 (0.70, 1.03)
Red meat
 Never or rarely24560.0 (53.8, 66.1) 1.0034.8 (29.1, 41.0) 1.00
 2/3 days per week159158.0 (55.6, 60.4) 0.92 (0.70, 1.21)25.1 (22.3, 27.3) 0.63 (0.47, 0.84)
 Most days188654.7 (52.4, 56.9)0.080.80 (0.61, 1.06)18.4 (16.7, 20.2)<0.0010.42 (0.31, 0.56)
Sports
 0 to 3 days per week181857.8 (55.5, 60.0) 1.0022.6 (20.7, 24.6) 1.00
 4 to 7 days per week194055.3 (53.1, 57.5)0.140.91 (0.80, 1.03)22.0 (20.2, 23.9)0.6740.97 (0.83, 1.13)
Watching television
 <1 hour per day30541.1 (35.6, 46.7) 1.0022.8 (18.4, 27.9) 1.00
 1 to <3 hours per day98054.4 (51.3, 57.5) 1.71 (1.32, 2.22)20.3 (17.9, 23.0) 0.86 (0.63, 1.17)
 3 to <5 hours per day103757.2 (54.2, 60.2) 1.92 (1.48, 2.49)22.1 (19.7, 24.7) 0.96 (0.70, 1.30)
 5 or more hours per day141460.3 (57.7, 62.8)<0.0012.17 (1.69, 2.80)23.2 (21.1, 25.5)0.4131.10 (0.92, 1.33)

Figure 2 shows the adjusted ORs of being overweight at age 14, comparing children of mothers who stated that they felt that family eating together was not important with those who felt that it was important and comparing mothers who reported that family had meal together a few times/once/less than once a week compared with those who had meal together at least once a day. These results are for the 2325 (61% of the 14-year-old respondents) children with complete data on all variables included in any of the multivariable models.

image

Figure 2. Adjusted ORs (N = 2325) of offspring being overweight at age 14, (A) comparing mothers who reported thinking that the family eating together was not important compared with those who thought it was important and (B) comparing mothers who reported that family had meal together few times/once/less than once a week compared with those who had meal together at least once a day. Model 1 was adjusted by age and sex at age 14. Model 2 was adjusted by age, sex, and being overweight at age 5. Model 3 was adjusted by age, sex, being overweight at age 5, parity, gross family income, maternal education, and race. Model 4 was adjusted by age, sex, being overweight at age 5, parity, gross family income, maternal education, race, maternal attitude toward the child's consumption of fatty foods and her report of the child's consumption of fast food, soft drinks, and red meat, and amount of television watching and participation in sports and exercise.

Download figure to PowerPoint

The age- and sex-adjusted associations (Model 1) do not differ substantively from the age-adjusted associations presented in Table 1 for respondents with complete data on each individual variable, suggesting that there is no substantive problem of selection bias within the subgroup with complete data on all variables. The offspring of women who felt that family eating together was not important had increased odds of being overweight at age 14 (OR, 1.27; 95% CI, 1.05, 1.53) in age- and sex-adjusted models. With additional adjustment for overweight status at age 5, this association increased slightly to 1.33 (1.08, 1.33). Further adjustment for other potential confounding factors (Models 3 and 4) did not substantially change this association. Maternal report of the frequency of family eating together was not associated with the child being overweight in any of the models. When we repeated the analyses using weights for factors that predicted non-response, the results did not differ from those presented here.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

In this study of Australian adolescents, we found that most mothers (78%) reported that the family ate together at least once a day, whereas just 43% felt that eating together as a family was important. The odds of being overweight at age 14 were greater among those whose mothers stated that it was not important that the family ate together. This association was not explained by potential confounding by a range of factors, including maternal education, family income, and maternal report of the child's consumption of fast foods and soft drinks and their frequency of television watching and participation in exercise and sports. However, maternal report of the actual frequency of family eating together was not associated with their child's overweight status. These findings suggest that maternal attitude towards family eating and diet may be more important than the actual frequency of family eating together in terms of offspring overweight. However, although not the focus of our main hypothesis here, it should be noted that the risk of being overweight at age 14 years was most strongly associated with overweight status at age 5 years.

Study Limitations

The participation at age 14 was 53%, and the participants who did not attend were more likely to be from poorer backgrounds, and mothers were more likely to have had low education and to be non-white. Our results would be biased only if the associations we have assessed were nonexistent or in the opposite direction in non-participants, that is to say, if among non-responders, those whose mothers stated that the family eating together was not important were less likely to be overweight. Although we cannot determine whether this is the case, we feel that it is unlikely. We have used a proxy measure (maternal report) for the true exposure of family eating patterns. However, direct observations of family eating patterns are not feasible in epidemiological studies. Due to the cross-sectional nature of our association of maternal perception of the importance of family meals and overweight status at 14 years, we cannot suggest causality of this relationship.

To assess whether those lost to follow-up produced bias in our results, we attached inverse probability weighting to subjects included in the analyses to restore the representation of those lost to follow-up. We followed the method suggested by Hogan et al. (22) and used robust SE estimates applied to the model. We found no difference between the weighted and non-weighted results (data not presented), which suggests that attrition is unlikely to have substantively biased our findings. We have compared our estimates of overweight or obese at ages 5 and 14 to the Australian National Nutritional Survey 1995 for similar age category, and the results are comparable. At age 5, the prevalence of overweight or obesity was 17% in MUSP and 15% in the National Nutritional Survey. Corresponding results at age 14 were 25% vs. 23%. The small differences are likely to be explained by regional variations.

Comparisons with Other Studies and Implications

If we consider that our exposure measure of maternal attitude towards the family eating together is a useful indicator of healthy family eating practices, then our findings are broadly consistent with previous reports that have found that children from families that eat regularly together are more likely to consume healthy diets (8, 9, 10, 11, 12, 13). There are several reasons why maternal attitude might be associated with the child's overweight status, but their actual report of the frequency of eating together is not. First, there could be measurement error in their report of eating together. To truly determine whether family eating patterns are associated with childhood overweight, one would ideally require direct observation of family eating patterns, which is difficult in large-scale epidemiological studies. Secondly, maternal attitude might encompass broader attitude towards the family diet nutrition and lifestyles and to that extent may actually be a more important measure of maternal effects on childhood obesity. This is demonstrated in Table 2, which shows the associations of fast food and soft drink consumption and frequency of television watching with maternal attitude toward family eating. However, adjustment for these characteristics did not substantively alter the association between maternal attitude to family eating and their offspring's overweight status at age 14. Thirdly, maternal attitude may reflect family eating patterns over a longer period of time than the time reported here; hence, those who stated that they felt that family eating together was important may have always eaten together frequently, whereas some of those who stated that they eat together now at least daily may have not done so in the past. Family eating patterns may influence childhood overweight patterns through a number of pathways. Having regular family meals together may reduce the amount of snacking in between meals, and families that eat together may have generally healthier diets. Eating together will enable the parent to have better knowledge of the child's food choices and amounts that they tend to eat. Finally, social interaction while eating may encourage children to perceive eating as a positive experience and establish positive attitude towards food, which continue into their later years.

In conclusion, we have found that adolescent offspring of mothers who reported that they felt that family eating together was not important are more likely to be overweight than those whose mothers reported that family eating together was important. These findings suggest that maternal attitude towards family eating and diet may be important determinants of childhood overweight status; however, replication of these findings in a longitudinal analysis is required to verify this finding.

Acknowledgement

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

We are grateful to all participants in the study. Greg Shuttlewood (University of Queensland, Australia) helped with data management for the study. The core study was funded by the National Health and Medical Research Council of Australia. This work was funded by the National Health and Medical Research Council (Grant 252834) and carried out at The University of Queensland and The Mater Hospital. D.A.L. is funded by a (United Kingdom) Department of Health Career Scientist Award. The views expressed in the paper are those of the authors and not necessarily those of any funding body. A.A.M. and D.A.L. developed the study aim and design. J.N., M.O., and G.W., set up and are responsible for the conceptual development and continued management of the Mater-University Study of Pregnancy and its outcomes. A.A.M. and D.A.L. undertook the analysis and wrote the first draft of the paper. All authors contributed to the final version of the paper.

Footnotes
  • 1

    Nonstandard abbreviations: MUSP, Mater-University study of pregnancy and its outcomes; CI, confidence interval; OR, odds ratio.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References