In the face of increasing health care costs, appropriate financing systems are not only required for somatic medicine but also for psychiatry. However, diagnosis-related groups (DRGs) did not lead to sufficient prediction of resource use for psychiatric inpatient care. Clinical case groupings besides groupings by diagnosis were rarely assessed so far.

We investigated whether the classical psychopathological syndromes according to the AMDP-System (Arbeitsgemeinschaft für Methodik und Dokumentation in der Psychiatrie) at admission would explain the variation of length of stay in psychiatric hospital. Using linear regressions with forward elimination, we analyzed the predictive value of the AMDP syndromes (n=998), and also of global assessment of function, illness severity and certain sociodemographic factors for the dependent variable length of stay (n=613).

Considering solely the AMDP-Syndromes within the linear regression resulted in a variance estimation of 6% with an influence of the paranoid-hallucinatoric and the psychoorganic syndrome on length of stay. Together with clinical and sociodemographic factors, we found a variance estimation of 20%, then with an influence of age, global assessment of function, severity of illness, voluntariness of admission and habiting situation, but no more the syndromes.

Our pilot study does not imply AMDP syndromes to be suitable for a case grouping to estimate length of stay. Other analytical approaches using AMDP might nevertheless lead to a better prediction, however, its use, however, shall be regarded critically.