We thank Professor Morgan for responding to our editorial and raising two important points. First, he is correct in saying that there was already some evidence favouring the outcome in anorexia nervosa if patients were enrolled in treatment as soon as possible after the onset of the illness. This came from his follow-up study of patients treated in Bristol where the emphasis was on local, easily accessible treatments. The outcome in the Bristol patients was significantly better than in those treated in two London hospitals providing ‘national services’ (Maudsley and St George’s Hospitals). This difference favouring Bristol was acknowledged by other experts in the field. ^{Reference Theander1} But in his 1982 article, Professor Morgan had already acknowledged the difficulty of assessing different therapeutic approaches in view of the selection of patients. It is inescapable that an evaluation of the treatment requires randomised controlled trials, as in the studies of family therapy reviewed in our editorial.

Professor Morgan’s second point was to stress that family processes are crucial in contributing to the success of early intervention in anorexia nervosa. He is right in recognising the risks of alienation in the patients’ relatives which undermines their contributions to a successful treatment. Again we welcome his observations enabling us to expand our too brief description of the essential principles of successful family therapy: (a) exonerating parents from causing the illness; and (b) getting them to take joint control of their child’s eating so that they are enabled to maintain a normal body weight.

These principles need some elaboration. Exonerating the parents requires the therapist to communicate a neutral position regarding the causes of the illness. The medical pioneers in this field of study (e.g. Gull, Charcot) expressed strongly negative views about relatives’ poor management of the problem, views which should be dispelled. Charcot’s influence was, of course, strongest in France where the cure d’isolement has only been abandoned within recent memory.

The second essential requirement is fraught with difficulties. Parents at first resist taking the necessary action. Their experience leads them to believe that they have failed to prevent their child’s poor eating and weight loss. Some parents fear that firmness on their part will lead to a loss of their child’s affection. They may also jump to the conclusion that an invitation to participate in treatment implies that they are being blamed. This can be combated by expressing the aims of therapy not as ‘changing the family’ but rather as helping them treat a sick family member. ^{Reference Russell, Gelder, Andreasen and Lopez-Ibor2}

Successful management requires an ongoing search for emotional and interpersonal factors (e.g. expressed emotion), which are responsible for maintaining (rather than causing) harmful behaviours. J.T. has contributed to a practical manual describing the techniques for negotiating successful transactions between carer and adolescent, focusing on rapport, language and problem-solving skills. ^{Reference Treasure, Smith and Crane3,Reference Goddard, Macdonald, Sepulveda, Nauman, Landau and Schmidt4}