| Abstract: | Despite renewed interest in drop‐out from eating disorders treatment, few studies have investigated the issue in respect to the most expensive and intensive form of treatment, that is, inpatient treatment for anorexia nervosa (AN). This study investigates whether risk of treatment drop‐out can be determined from information routinely collected at point of admission. Using information from a multi‐site database collected in Australia and New Zealand, demographic and clinical data at point of admission were collated for 213 inpatient treatment episodes. One in five admissions ended with the patient unilaterally deciding to leave treatment without clinician endorsement. A lower body mass index, AN purging subtype and active fluid restriction made significant independent contributions to this risk. Drop‐out remains a highly disruptive method of discharge and while there is utility in predicting those most at risk, few variables commonly collated by clinicians contribute to their identification. The implications for clinical practice and future research are discussed. Copyright © 2003 John Wiley & Sons, Ltd and Eating Disorders Association. |
