How to Help Someone With an Eating Disorder

If you have just worked out that someone you love has an eating disorder, three things matter before anything else.

First, know what counts as an emergency. The Royal College of Psychiatrists’ guidance for relatives lists the signs that call for urgent assessment:

• fainting when standing up;
• being unable to climb stairs;
• talking about ending their life;
• suddenly refusing food or water;
• another illness, such as diarrhoea or vomiting, at the same time.

If you see any of them, it says, encourage the person to go to the emergency department or contact their eating disorder team immediately.

The same guidance gives clinicians further signs of high risk to life: acute confusion, refusing fluids, and being unable to sit up from lying flat or get up from a squat without using the arms. It adds that chest pain, palpitations and breathlessness can be signs that the heart is struggling, and that low calcium can lead to seizures. In an emergency, call 999 in the UK or 911 in the US.

The guidance warns that a person can be at high risk while appearing well and having normal blood tests. Normal results are not a reason for reassurance. Some people appear very energetic right up to a physical collapse, and sudden cardiac death does occur in anorexia.

Second, do not start a confrontation about food. Beat advises a place where you both feel safe and will not be disturbed, a time when neither of you is angry or upset, and a conversation not centred on food or weight. NEDA, the US National Eating Disorders Association, advises “I” statements about behaviour you have seen yourself, because sounding accusatory can make someone defensive.

Third, say the two things that help: that it is not their fault, and that people recover. Beat’s tips for supporters include both. Do not reassure them about their body; the section on what to say explains why.

Beat’s general tips for supporters open with two lines: recognise that you are not to blame, and acknowledge to your loved one that they are not to blame.

That advice comes after a long history of blaming families. Some nineteenth-century doctors treated families as part of the problem. In 1874 William Gull advised that patients be surrounded by people with moral control over them, “relations and friends being generally the worst attendants”. In the 1880s Jean-Martin Charcot made separating patients from their parents a condition of his treatment.

In 1978 Salvador Minuchin and colleagues described enmeshment, rigidity, over-protectiveness and conflict avoidance as typical of families of people with anorexia.

In 2010 the Academy for Eating Disorders, a professional association, said that research clearly refutes the notion that families are a primary cause, and it condemns statements that blame families. Its position is narrower than it sounds, though: it says particular styles of family behaviour and biological vulnerability may raise the risk of mental health problems, eating disorders included.

Family-based treatment, the family therapy most trials now test, is explicitly agnostic about cause and treats parents as the main resource for recovery.

Families may still blame themselves. The national guideline for England tells clinicians to be aware that family members may feel guilty and responsible for the eating disorder.

Charities for carers list sentences to say and to avoid.

The National Alliance for Eating Disorders suggests “I might not understand, but I love you and am here to listen.”

F.E.A.S.T. is an organisation for parents and caregivers. At a difficult meal, a phrase book on its website suggests short lines such as “I love you and I will keep you safe” and “I hear how hard this is for you. You can do hard things. I’m here”.

Any version of just eat. The National Alliance lists “You just need to eat more/less” among things not to say. Beat explains why: eating disorders are not a choice but severe mental illnesses.

Beat’s list of remarks people may make innocently shows how the illness can twist them. “Just eat normally” may be heard as “You’re not trying hard enough” and “it’s your fault”. The National Alliance also warns against “You’re overreacting.”

And anything about how they look. Including the kind things.

Both charities give the reason. Beat says “You look well” may be heard as “you have gained weight”, and that comments about looking “healthier” or “better” are often taken to mean weight gain. The National Alliance says such comments keep attention on the body and can trigger the eating disorder, even when they are meant kindly.

Clinicians often separate the illness from the person, talking about what the anorexia wants rather than what the patient wants. The practice is called externalisation, and it is a key part of family-based treatment. Beat suggests families try it, but warns that some people feel the eating disorder is part of them and find such questions patronising.

A 2024 study interviewed 13 adults in recovery from anorexia. Most said that other people’s externalising language early in recovery could cause conflict, disconnection and distance with family and close others. Externalisation helped most when the person led it in their own words, and least when it did not let them feel seen as a person beyond the illness.

Reorganising your household around someone’s illness is called accommodation, and there is a scale that measures it: the Accommodation and Enabling Scale for Eating Disorders.

It looks like cooking a separate meal. Buying only the safe foods. Letting rules about which plate, which cutlery, which time of day quietly govern what everyone else does. Answering the same anxious question again, and again. Not booking the restaurant. Not going on the holiday. Seeing something and deciding not to mention it.

In the study that built the scale, carers who spent more than 21 hours a week with the person accommodated more. The scale’s authors describe carers becoming increasingly entrapped in the rules of the illness.

Whether accommodation harms the person is less clear. In one small study of 40 relatives of patients in residential treatment, more accommodation went with worse eating disorder scores at discharge in the 18 patients tracked. Accommodation fell slightly after a skills programme in the study that built the scale, which had no comparison group. In the ECHO trial of a carer course, it did not differ between the groups. In a 2017 pilot trial of the same course for parents of adolescents, it was slightly lower at six months.

None of these studies shows that reducing accommodation helps the person recover. The scale’s authors found that accommodation went with difficult behaviours such as threats and self-harm, not with how severe the illness was.

If you are supporting someone through meals, NHS Lothian and Beat publish practical guidance, and it largely agrees.

Agree the plan beforehand so that nothing is negotiated at the table. Have everything ready before you sit down: Beat says last-minute changes can cause panic and lead people to eat less. Keep the conversation light and away from food, weight and bodies. Eat the same meal alongside them, as NHS Lothian’s guide for parents and carers suggests.

Then stay.

Support does not end with the meal. NHS Lothian’s guide says the person may feel extremely anxious and guilty after eating, and suggests distraction such as card games, crafts, a drive or a programme together. It also asks carers to stay calm and to be consistent.

Some of the advice above, such as family-based treatment, is built around parents and does not transfer to other relationships.

If the person is male, Beat warns that eating disorders are widely seen as feminine. Men may be less likely to see themselves as having one, and others, healthcare professionals included, may be less likely to suspect it. Beat also says men face more stigma.

Lehe and colleagues reported two surveys, in 2024 and 2025. In both, men who saw more stigma around eating disorders in men reported less intention to seek help. The link held only for symptoms seen as feminine, such as concern with thinness. The 2025 survey found no such link in women.

In a study from Spain, the burden of caring was highest for eating disorders. It compared 251 carers of people with eating disorders in Spain with carers of people with depression or schizophrenia in other European countries, using the same questionnaire. Being a mother or a partner predicted a heavy burden.

In a sample used to test the accommodation scale, 61% of carers scored at or above the clinical threshold for anxiety before a skills course.

Courses for carers have mixed results: they have helped in smaller studies, but the largest trial found no benefit to carers or patients. A 2015 review of 13 studies found that most courses produced a moderate drop in carer distress. Most courses also produced small to moderate drops in burden and in expressed emotion, a measure of criticism and over-involvement in the family, and the changes lasted.

The ECHO trial, published in 2015, recruited patients with anorexia and their carers from 15 UK inpatient services. Its 178 patients and 268 carers were randomised, by family, to a carer skills package or usual care. Carers given the package had less burden and expressed emotion, and spent less time caregiving, at six months, but these effects had diminished by twelve. The patients in those families used fewer inpatient bed days seven to twelve months after discharge.

At two years, the ECHO group was ahead on every outcome, for patients and carers, by small to moderate amounts. But Eating Disorders Review, summarising the two-year paper, reports that none of the differences between the two groups was statistically significant.

The largest trial since, TRIANGLE, was published in 2024. It randomised 371 patients with anorexia, together with one carer each, to an online self-help programme for patients and carers or to usual care. The patients were aged 16 or over and leaving intensive treatment at 31 UK services. The trial found no significant difference in carers’ depression, anxiety and stress, or in their caregiving skills, at 12 or 18 months, and none in the patients’ outcomes.

Only 20% of patients and families used enough of the materials to meet the trial’s target.

A Spanish trial that randomly allocated 63 carers found a skills workshop no better than a psychoeducation workshop, though carers in both groups improved.

Someone can be seriously ill and decline all of it. Beat notes that denial is common, and that in anorexia it is considered a symptom of the illness, so a refusal is not a failure of your approach.

What you can do: Beat suggests making a GP (family doctor) appointment for them and offering to go with them. That way you can speak on their behalf if they would find it helpful. If they refuse help, Beat advises trying to raise the issue again as soon as possible. If they are physically unsafe, go to emergency care. Tell the doctors what you have seen: the Royal College of Psychiatrists’ guidance tells clinicians to take parents’ and carers’ information into account when assessing risk.

What you cannot do is detain anyone on your own. In England and Wales, and in Scotland, an eating disorder counts as a mental disorder under mental health law. The courts in England and Wales have held that feeding is a core part of treatment for anorexia. So feeding can be given compulsorily under that law, there and in Scotland, at any age.

In England and Wales detention needs the recommendations of two doctors. The application is made by an approved mental health professional or by the patient’s nearest relative, as the Mental Health Act defines that person.

In the United States, commitment is ordered by a judge, or someone acting as one, under state laws that vary. Danger to self or others is the usual test. Nearly every state also allows other grounds, such as “grave disability”, typically an inability to provide for basic needs for food, clothing or shelter.

Scotland operates under separate legislation with an additional impaired-decision-making test, thresholds and procedures vary considerably between US states, and the position for a minor differs again. Anyone facing this needs advice on their own jurisdiction.

Suicide is a real risk in this illness. A 2011 review of 40 studies found that people with anorexia died by suicide at 31 times the rate expected in the general population. However, the risk fell in recent decades.

Asking someone directly whether they are having suicidal thoughts has not been found to raise their suicidal thinking. A review of studies in research and clinical settings found no significant increase after people were asked. Samaritans advises asking: it says evidence shows that asking someone if they are suicidal could protect them.

Then tell their treatment team, and help keep the environment safe. The national guideline for England on self-harm says a safety plan should be developed in collaboration with family and carers, as appropriate.

Expressed emotion is a measure of the family environment, covering criticism and emotional over-involvement. It came from schizophrenia research, where high expressed emotion reliably predicts relapse. A 1998 review found that it predicted outcomes in eating disorders even more strongly than in schizophrenia. But criticism and emotional over-involvement may not behave alike in anorexia.

One study in France followed 60 adolescent girls with anorexia. High emotional over-involvement in mothers was associated with better outcomes 18 months later, after allowing for treatment and for how ill the girls were at the start. The girls had better nutritional status, lower eating disorder severity and fewer rehospitalisations. No link was found for fathers.

Criticism, by contrast, has been shown to have a negative influence on treatment outcome, the accommodation scale’s authors note. Yet carer trials such as ECHO count both parts in the expressed emotion they measure.

People who were treated against their will often see it differently afterwards, depending on how they are doing. In a study of 30 adults who had been treated involuntarily for anorexia, those now doing well saw the treatment as beneficial in retrospect. Those still struggling reported negative consequences.

In an earlier study, 46 of 139 patients admitted to a Baltimore eating disorders programme did not at first agree that they needed admission. Two weeks later, 20 of them had changed their minds. In an Oxford interview study of 29 young women, compulsory treatment was considered appropriate where the illness was life-threatening.

The same intervention is remembered two opposite ways, and which way appears to depend on how things turned out. If you are weighing this decision for someone else, the evidence does not tell you what to do.

Most of what families are told is guidance from charities, national guidelines and carer researchers, not the result of a trial. That covers what to say and how to support meals. The Royal College of Psychiatrists warns that a person can be at high risk while appearing well and having normal blood tests.

Where help for families has been tested, the results are mixed. In a 2015 review of 13 studies, most courses for carers produced a moderate drop in carer distress, but the largest trial, TRIANGLE, found no benefit to carers or patients. The studies of family accommodation described above do not show that reducing it helps the person recover.

Two findings complicate the advice. In one study of adolescent girls with anorexia, high emotional over-involvement in mothers went with better outcomes. And people treated against their will often judge the treatment afterwards by how they are doing.

Download the conclusion

The verdict above as one image, with the number of patients in two trials of courses for carers. ECHO had 178, and carers’ gains at six months had diminished by twelve. TRIANGLE had 371, and found no significant difference for carers or patients. It is 1,200 by 630 pixels, and anyone may download it and post it.

Download the card (PNG, 1,200 × 630)