Any parent who has had a stand-off with a toddler about whether they’ll eat their tea has had a small taste of the anxiety of not being able to feed your child.

How must it feel to watch an eating disorder take over your daughter (or son)? Every mealtime an uphill struggle to help them fight the illness by getting some nourishment into them. Trying to help them stop making themselves sick, secretly over-exercising, abusing laxatives, binging. Worrying about the effects on their health, on their school-work or job, on their friendships, on their self-esteem, on the rest of their life. Not recognizing the child they used to know so well. Not knowing how best to help them.

Working with parents of young people with eating disorders I see the immense and exhausting pressure that these illnesses put on parents and carers.

Parents or carers often play a central role in supporting their child’s recovery; but what is the evidence on how professionals can support them? What works, and how should it be delivered? This is the question Hibbs et al set out to answer in their meta-analysis of interventions for care-givers of someone with an eating disorder.

Methods

The authors conducted a meta-analysis of the published literature on interventions for carers of people with eating disorders. They followed PRISMA guidelines and used appropriate searches and analysis techniques. Although unpublished literature was not sought, the authors did test for publication bias.

The hypotheses were that carer outcomes would improve post intervention, and that improvement would be greater with more intensive interventions.

Results

The authors selected 13 studies, but only 5 were randomised controlled trials (RCTs). They were largely conducted in the UK, and ranged in size from 13 to 268 participants. The mean age of patients they were caring for ranged from 13 to 26 years old.

Interventions varied and included:

• Workshops
• Self-help online resources/books
• Phone/online/face to face support

The authors looked at pre- and post-treatment scores on several measures:

• Carer Distress
• Carer Burden
• Carer Contentedness
• Carer Expressed Emotion

They did not conduct any meta-analysis on the differences between treatment and control groups.

They found moderate effect sizes for reduction in Carer Distress, and small to moderate effect sizes for reduction in Carer Burden and Carer Expressed Emotion following interventions and at follow-up.

Discussion

This is a good, thorough meta-analysis; the searches and analysis were appropriate and followed PRISMA guidelines. There are obvious problems with the studies testing different patient populations, and different interventions, making it hard to directly compare them, and it is a shame that any non-English language studies would have been missed.

In my view the main problem with the study is that the meta-analysis is of changes in scores before and after the interventions. Yes, people seem to get better, but interpreting this is plagued with problems:

• People might get better over time anyway, even if no-one intervenes
• Carers might get better because their children (who are being treated) get better
• Just being in a trial and a lovely Research Assistant taking the time to ask how you are might help you feel better
• You might really like and want to please your therapist so you say you feel better at the end to please them
• There is also a statistical problem called “regression towards the mean”, which says that in mathematical terms, if you take repeated measurements of the same sample, and your first measurement is extremely high, then your next measurement is likely to be lower (nearer to the average or “mean”), and vice versa

Randomised controlled trials are designed to overcome these problems, making sure that any improvements we see are related to the treatment and not changes that have occurred by chance or for other reasons. They compare a treatment either to treatment as usual/waiting list control, or (more stringently) to an attention control (where the control group gets an equal amount of attention from equivalent professionals but without whatever is viewed as the “active ingredient” given to the treatment group).

Sadly, a good meta-analysis of limited studies doesn’t generate good evidence. This meta-analysis mainly demonstrates the shortage of RCTs in this field (although, tantalizingly, it references the author’s own, as yet unpublished RCT).

We need rigorously designed and conducted RCTs, ideally comparing interventions to active controls, so we can find out what works. In the meantime, we are left with common sense advice:

• Clinicians: keep supporting carers in the best ways you know
• Parents and carers: look after yourselves and ask for help when you need it

Link

Hibbs, R., Rhind, C., Leppanen, J. and Treasure, J. (2014), Interventions for caregivers of someone with an eating disorder: A meta-analysis. Int. J. Eat. Disord.. doi: 10.1002/eat.22298 [PubMed abstract]

Carers’ Stories. Beat website, last accessed 1 Dec 2014.