Limited predictive value of admission time in clinical psychiatry

Abstract

Background A large proportion of admissions to psychiatric hospitals happen as emergency admissions and many of them occur out of core working hours (during the weekends, on public holidays and during night time). However, very little is known about what determines admission times and whether the information of admission time bears any relevance for the clinical course of the patients. In other words, do admission times correlate with diagnostic groups? Can accumulations of crises be detected regarding circadian or weekly rhythms? Can any differences between workdays and weekends/public holidays be detected? May it even be possible to use information on admission times as a predictor for clinical relevance and severity of the presented condition measured by the length of stay? Methods In the present manuscript we analyzed data derived from 37′705 admissions to the Psychiatric District Hospital of Regensburg located in the Southern part of Germany covering the years 2013 to 2018 with regard to ICD-10 diagnostic groups and admission times. The hospital provides 475 beds for in-patient treatment in all fields of clinical psychiatry including geriatrics and addiction medicine. Results Several core questions could be answered based on our analysis: 1st Our analysis confirms that there is a high percentage of unheralded admissions out of core time showing broad variation. 2nd In contrary to many psychiatrists’ misconceptions the time of admission has no relevant impact on the length of stay in the hospital. 3rd The predictive value of admission time regarding the allocation to ICD-10 diagnostic groups is low explaining only 1% of variability. Conclusions Taken together, our data reveal the enormous variation of admission times of psychiatric patients accounting for the need of adequate and consistent provision of personnel and spatial resources.