Inpatient long-term video-EEG monitoring (LTM) is an important diagnostic tool for patients with seizures and other paroxysmal behavioural events. The main referral categories are diagnosis (epileptic versus non-epileptic disorder), seizure classification and presurgical evaluation. The diagnostic usefulness of the LTM varies considerably (19-75%) depending on how this was defined and on the selection of the patients. The purpose of this study was to assess the diagnostic usefulness and the necessary duration of the LTM for the referral groups, in patients extensively investigated before the monitoring. An LTM was considered diagnostically useful when it provided previously not reported, clinically relevant information on the paroxysmal event. For the presurgical group, reaching a decision concerning surgery was an additional requirement. We reviewed data from 234 consecutive LTM-sessions (221 patients) over a 2-year period. In 44% of the cases the LTM was diagnostically useful. There were no significant differences concerning diagnostic usefulness among the main referral groups: diagnostic (41%), classification (41%) and presurgical (55%). Diagnostic usefulness did not differ among the age groups either. The duration of the successful LTM-sessions was significantly longer in the presurgical group (mean: 3.5 days) than in the diagnostic and classification groups (2.4 and 2.3 days, respectively). We conclude that LTM is a valuable diagnostic tool even in patients extensively investigated before the monitoring, and is equally effective in the referral and age groups. However, patients referred for presurgical evaluation need considerably longer LTM, and this should be taken into account when planning the resources and calculating the costs.