This paper reports on a case in which family therapy was one of the approaches attempting to resolve an eating disorder in the second‐born of twins, the youngest in an intact family of four children. The case is described in its referral, agency and chronological contexts. The initial referral suggested that there was no physical cause for the presenting problem, and advocated an investigation of family dynamics. Some evidence which seemed to support the dysfunctional family formulation was provided, and various hypotheses regarding family disturbance were pursued in therapy. These hypotheses were, on the surface, supported by systemic observations. When a rare pancreatic tumour which was underlying the eating disorder was diagnosed many months after the brief and unsuccessful interventions, ‘confirmation bias’, resistance and neutrality became important issues for the therapists to consider. The role of these processes in this case is disc