Factors that predict psychiatric consultation in the emergency department

The purpose of the study was to determine the basis for nonpsychiatrists' decisions to obtain psychiatric consultations in an emergency department (ED) for patients with transient to moderate psychiatric symptoms. The authors reviewed ED records during July-August 1982 and January-March 1983. Complete data were gathered for 133 of 134 patients with primary or secondary psychiatric discharge diagnoses and scores on the Global Assessment Scale (GAS) such that most clinicians would not think immediate consultation was required. Information included five sociodemographic variables, psychiatric history, GAS scores, type of interaction with staff, and specificity of medical complaints. Interrater reliability of GAS coding was 81%. A total of 50% of the patients received psychiatric consultations in the ED. Both psychiatric history and GAS scores were associated with consultations (P less than 0.001 for both). Using stepwise logistic regression, psychiatric history was more important in predicting consultation (SRR = 3.0, P = 0.001) than was GAS score, which did not enter the model as a significant predictor. Social characteristics also distinguished between patients who received a psychiatric consultation in the ED and patients who did not. It was concluded that nonpsychiatrists base important management decisions on a history of psychiatric treatment for patients with transient to moderate symptoms. Whether this is appropriate needs to be studied.