Comments by Kellie Lavender and Dr Roger Mysliwiec from NZEDC regarding the recent Meta Analysis paper: Treatment outcomes for anorexia nervosa: a systematic review and meta-analysis of randomised controlled trials (2018). Murray, S., Quintana.D., Loeb, K., Griffiths, S., Le Grange, D.

We hold some concerns about the message of the paper, which has arrived at the conclusion that current specialised treatments have no more advantages over comparator treatments as usual in terms of psychological symptoms and no advantage in terms of weight recovery at follow up.

We are concerned that the message this paper inadvertently sends could do significant harm to patients and their families and to clinicians and we believe that the message to this paper needs to be refined.

The conclusion that as a whole many interventions in the treatment of anorexia are not effective is already well established and not new and especially applicable to the treatment of adult anorexia. It is also not surprising that the paper reached that conclusion given that the meta-analysis included studies of interventions that we already know do not work effectively, including pharmacological interventions, hospitalisations and many adult treatment interventions.

The paper appears to have some fundamental methodological problems, which are worth noting.

In order for a meta analysis to come to reliable conclusions it is essential to include enough well designed and comparable studies. In this review a lot of studies are included comparing different interventions and outcomes, in different contexts and for different populations.

By including studies with interventions that we already know do not work with studies that have shown interventions to work (like the FBT vs. AFT trial in 2010), the potency of a specific intervention gets diminished and watered down.

There are also a number of more specific issues with the studies that have been included that need to be considered and usually would have precluded them from a meta-analysis. Some studies included were only feasibility studies, which by definition are not designed to test effectiveness of treatment. Even though some of the studies have been randomised they have been very underpowered and would therefore not be considered a well designed RCT, which would usually require an n >60).

In many of the studies included outcomes were not measured well and timelines were too varied (EOT varied from 7 weeks to 12 months and additionally the endpoints were measured differently).
The studies with the largest numbers were studies comparing similar type of interventions rather than comparing specialist treatment with treatment as usual, which the meta analysis had originally intended to do.

A major point is also that the follow up data were based on very small limited data of only a small number of the included EOT studies.

This means that the conclusions of this paper are not based on reliable data and sound methodology. Apples are not being compared with apples. An illustrative metaphor would be if one was taking a cookbook, include all the ingredients in the book in one recipe – with the outcome that the prepared food was not tasty and then arriving at the conclusion that every recipe in the cookbook was terrible.

As an aside to the general issues of the meta-analysis itself, we would also like to comment on the paper’s reference to describing family interventions as focusing on weight recovery only. When manualised Family Based Treatment is done with fidelity there is a sustained expectation of weight recovery, as well as a focus on systematic extinction of Eating Disorder behaviours, and helping parents manage the inevitable distress that the young person will experience. Most importantly, Family Based Treatment studies do show full recovery of weight and psychological symptoms at EOT and follow-up at rates between 30 to 45%.

We think that the message of this paper is problematic as patients and affected families might conclude that there is no point (and no hope) in engaging in specialist treatments and therapists and service providers might use this as a rationale that it does not matter whether clinicians have training in specialist treatment nor that it is important to ensure treatment fidelity.

The conclusion that we need to keep finding new effective treatment is of course very true. However, we do have strong evidence for the effectiveness of Family Based Treatment (FBT) to treat anorexia in the adolescent population and this should be highlighted. To send a message that nothing works is not only unhelpful but also incorrect. We feel it is important that families, who are fighting for the recovery of their child, know that there is evidence for effective specialist treatment and that by intervening promptly and effectively there is a realistic hope for full recovery.

The New Zealand Eating Disorder Clinic is a private specialist outpatient clinic in Auckland, New Zealand, providing evidence based specialist treatment for eating disorders. Kellie Lavender is a specialist in Family Based Treatment and a faculty member of the Training Institute for Child and Adolescent Eating Disorders.

Dr Roger Mysliwiec is a medical doctor, specialist in Psychosomatic Medicine and Psychotherapy (Germany). Roger has over 30 years of experience working with eating disorders and is considered one of New Zealand’s leading experts in the field of eating disorder treatment.