OBJECTIVES: To determine whether loss to follow-up can be predicted in patients who present to an emergency sexual assault assessment service and to generate hypotheses regarding the prediction of loss to follow-up on the basis of patient characteristics, assault characteristics and the services provided. DESIGN: Prospective, exploratory study. SETTING: Emergency department functioning as a regional sexual assault centre in a tertiary care hospital. PATIENTS: All 294 women over the age of 16 years who presented to the emergency department with a complaint of sexual assault and consented to be followed up. INTERVENTIONS: Telephone interviews at 24 to 48 hours and 1 month after presentation; face-to-face interviews after 1 week, 3 months and 6 months. MAIN OUTCOME MEASURES: Follow-up status (tracked versus lost to follow-up), State-Trait Anxiety Inventory (STAI-Y), Beck Depression Scale (Beck) and Rape Trauma Symptom Rating Scale (RTSRS). RESULTS: At 24 to 48 hours 136 (46%) of the patients could not be reached. Only 61 (21%) were still tracked at 6 months. Loss to follow-up at 1 month accurately predicted loss to follow-up at 6 months in 209 (98%) of 214 patients. For tracked patients the STAI-Y and Beck scores improved over 6 months. These scores at 1 week did not predict follow-up status at 6 months, but the numbers were small. Subjects with a higher RTSRS score at 24 to 48 hours were most likely to remain tracked throughout the 6 months. CONCLUSIONS: Decisions regarding how vigorously to track patients with a complaint of sexual assault can tentatively be based on the characteristics of the victim and of the assault. We hypothesize that the characteristics predicting loss to follow-up include denial and avoidance behaviour, lack of a telephone number or forwarding address, history of a psychiatric condition, a disability (e.g., deafness), characterization as a "street person," a high degree of violence or injury in the assault, and threat by the assailant. Although a predictive model requires further data, crisis intervention services in an emergency department are essential, given the large number of patients lost to follow-up.