I declare a vested interest. I grew up eating regularly with my family, not seeing it as anything special, just what we did: sometimes a chance to talk about the day, sometimes a time to debate politics and social justice (like whether you or your sister has been given more peas). Now I’m a mother myself, sitting down to eat with my husband and son is one of my favourite things.

Clinically, eating together as a family is something I advocate in treating eating disorders in children and adolescents; I would also suggest it to parents worried that their child might be developing an eating disorder.

Let’s see if this systematic review of the literature supports my gut feeling that eating together can be a positive thing in terms of psychosocial outcomes.

Methods

The authors stated aim was to “conduct a systematic review of the effects of family meals on psychosocial outcomes in children and adolescents”. Studies were identified by searching the MEDLINE and PsycInfo databases, and by reviewing bibliographies of relevant papers. To meet the inclusion criteria, studies had to:

1. Be published in a peer-reviewed journal in English
2. Discuss the role of family meals on psychosocial outcomes (defined as: substance use, disordered eating, depression) in children/adolescents
3. Be cross-sectional, longitudinal cohort or randomised controlled trials

Two authors independently assessed the articles.

Results

Of 1,688 articles screened, only 14 met the inclusion criteria, of which 7 were longitudinal studies.

• Five studies did not define what constituted “frequent family meals”; the majority defined this as at least 5 per week.
• 33-61% of children in the studies reported “frequent family meals”, in Britain this was only 33%.
• Nine studies explored disordered eating, with most finding that family meals are protective against disordered eating, after adjusting for plausible confounders of the relationship like “family connectedness”, socioeconomic status and personal characteristics.
• Interestingly the results only hold for females. I wonder if that is because these types of behaviours are less common in males, so the studies may have been underpowered to detect an association.
• Family meals also seem to be protective against body dissatisfaction and drive for thinness.
• They also explored substance use, finding family meals to be associated with reduced smoking of both tobacco and cannabis, alcohol and illicit drug use. Again, results were more consistent for girls.
• Other outcomes were largely based on results from a single study.

Discussion

• So, this is all OK in terms of gathering the literature, though it’s a shame they didn’t track down unpublished or translate non-English studies.
• They point out the obvious problem with cross-sectional studies (they tell you nothing about direction of causality) and try to make sure longitudinal studies adjust for confounders (lower socioeconomic status is associated with fewer family meals).
• It’s hard to do this comprehensively (how can you totally capture the type of family who eats together?) but I think it would be relevant, for example, to control for parental history of an eating disorder (we know this is associated with eating disorders in offspring (shameless self-citation: Bould et al., 2015), and a parent with such a history might well find family meal times stressful and try to avoid them.

The paper includes an excellent table describing the different studies, sadly their table of results includes a pet hate of mine; referring to results as “statistically significant” without giving us any actual numbers to look at. This lack of point estimates, confidence intervals and effect sizes makes it very hard to interpret apparent sex differences, and to have any sense of the magnitude of the effect of family meals.

I am a reasonably good Elf, so I read one of the original studies for you (Neumark-Sztainer et al, 2008) (I’m not such a good Elf as to actually rewrite the entire paper). This is a nice longitudinal study:

• They do indeed find that, for girls, family meals are protective against “extreme weight control behaviours”, such as vomiting, laxative/diuretic/diet pill use (adjusted odds ratio 0.71 (0.52 to 0.97), p=0.03 in case you’re a stats junky too – that’s reducing the odds of extreme weight control by a third).
• When you look at the numbers (adjusted OR 0.8 (0.45 to 1.40), p=0.43), the lack of apparent association for boys may indeed be because of a lack of power:
  • 22% (301/1,363) of girls had extreme weight control behaviours
  • Compared to 7% (74/1,115 ) of boys.
• Unexpectedly, they also found that in boys, frequent family meals at Time 1 actually increased the risk of skipping meals (OR 1.81, 95% CI 1.24 to 2.63) and eating little food at Time 2. I don’t think this is about power, but like the authors, I’m at a loss to understand why it is (space for your suggestions below).
• It’s worth noting that they didn’t adjust for family history of eating disorders.

Summary

In summary, the lack of numbers and detail in this systematic review leaves me feeling rather under-informed. On the whole, family meals seem like a good idea, especially for girls (and I’m not going to stop even though I do have a boy), but to really answer this question, at the very least we need a better systematic review.

Links

Primary paper

Harrison ME et al. Systematic review of the effects of family meal frequency on psychosocial outcomes in youth. Can Fam Physician 2015 Feb;61(2):e96-e106.

Other references

Bould, H., Sovio, U., Koupil, I., Dalman, C., Micali, N., Lewis, G., & Magnusson, C. (2015). Do eating disorders in parents predict eating disorders in children? Evidence from a Swedish cohort. Acta Psychiatr Scand. doi: 10.1111/acps.12389 [Abstract]

Neumark-Sztainer, D., Eisenberg, M. E., Fulkerson, J. A., Story, M., & Larson, N. I. (2008). Family meals and disordered eating in adolescents: longitudinal findings from project EAT. Arch Pediatr Adolesc Med, 162(1), 17-22. doi: 10.1001/archpediatrics.2007.9