Predicting the unpredictable: issues for output-based funding in psychiatric services.

In this study a number of important issues relating to diagnosis related groups and psychiatry are highlighted. Retrospective analysis of admissions to a rural psychiatric admission unit was undertaken utilising two differing approaches to statistical analysis. Results revealed that diagnosis alone is not highly predictive of cost. However, when linked to other variables, greater explanation in variation of length of stay was achieved. Data from social variables yielded sufficiently interesting results to warrant further investigation. The authors conclude by reporting the resultant groups from the classification analysis and offer some suggestions for further research.