A23 23..23

Three papers in the Journal this month examine the impact on service use of particular service developments and patient/illness characteristics. Williams et al (pp. 480–485) compared patterns of service use, including hospital admission rates, before and after the introduction of a triage admission system in one large innercity mental health organisation. Overall, little difference was found between the two conditions – in terms of average length of total hospital stay, number of readmissions or cost – even though length of stay on the actual triage wards was reduced for many patients. The authors comment on the possibility that transfer from triage wards for those who cannot be rapidly discharged may actually prolong admission for that group. Focusing on patient/illness characteristics, Sara et al (pp. 448–453) utilised mental health service use data to examine the impact of cannabis and stimulant disorders on outcomes for those with first-episode psychosis. Over the 2-year follow-up period, readmission was not found to be predicted by the presence of baseline substance use disorders, although both the presence of a stimulant disorder diagnosis made prior to the index admission with psychosis and ongoing problem drug use did predict readmission. The authors comment on the need to identify substance use disorders for those with first-episode psychosis since intervention leading to discontinuation may be associated with the best outcomes for this group. D’Amico et al (pp. 441–447) undertook a long-term follow-up study of service use and associated costs for a sample of young adults with a childhood history of hyperactivity/conduct problems. The authors found an association between high levels of childhood conduct problems and early adulthood service costs, particularly driven by involvement with the criminal justice system. Interestingly, across all baseline groups in the sample, high levels of A&E and general hospital contact were found in young adulthood while use of mental health services was comparatively lower.