Clinical predictors of readmission to psychiatric inpatient care: A 20-year follow up study of former adolescent inpatients

BACKGROUND: Readmission is a poor outcome associated with significant economic and psychosocial burden, particularly among young people. This study aimed to investigate predictors of time to readmission, the number of readmissions, and the cumulative duration of all readmissions among adolescent psychiatric inpatients over a 20-year observation period. METHODS: A total of 508 adolescents participated in the original study. The length of index hospitalisation, previous inpatient admissions, the number of psychiatric diagnoses, and the severity of depression, anxiety, mania, psychosis, obsessive-compulsive disorder, eating disorder, conduct disorder, and alcohol and substance use were used to predict readmissions. Separate Bayesian regressions were conducted to examine the impact of these predictors on time to readmission, number of readmissions and the cumulative duration of readmissions. RESULTS: The severity of psychosis symptoms predicted all three outcomes. Once participants with existing schizophrenia spectrum diagnoses were removed from the sample, psychosis symptoms still predicted time to readmission and the number of readmissions. Previous inpatient admissions predicted more frequent admissions during the observation period. The severity of depression symptoms was associated with shorter time to first readmission. DISCUSSION: Looking at a range of patient and service level measures, psychosis symptoms predicted all three readmission outcomes in psychiatric inpatients. This finding suggests that psychosis symptoms may be a useful transdiagnostic marker of illness severity, predicting poor outcome into adulthood.