Bedside cognitive screening tests have been suggested as a way to quickly determine the likelihood of a patient's having an organic mental disorder (OMD). The Mini-Mental State (MMS) examination and a number of separate bedside neurobehavioral tests were administered to 206 patients who were consecutively admitted to the University of Iowa Psychiatric Hospital. Both examinations were completed by 150 subjects. Of the sample, 7% had an OMD diagnosis. The most frequent neurological causes of OMD were probable Alzheimer's and closed head trauma. The predictive value positive, predictive value negative, specificity, and sensitivity of the MMS were examined for a number of different cutoff scores. The maximum sensitivity was 75% with a cutoff score of 30. The maximum specificity was 100% with a cutoff of 12. However, only two OMD patients had scores less than 12. There was no cutoff that resulted in a clinically useful screening test. A stepwise multiple logistic regression model using the separately administered bedside tests as independent variables and OMD as the dependent variable resulted in two significant predictors of an OMD diagnosis: the alternating sequence test and a reading test. This model was not appreciably better than the MMS in predicting an OMD diagnosis. A number of problems make the development of a clinically useful cognitive screening test for OMD in inpatients unlikely. These include poor cooperation of the patients, the rarity of OMD in psychiatric inpatients, the heterogeneity of OMD etiologies, and the confounding neuropsychological deficits of a number of common psychiatric disorders.