Objective analysis of olfactory identification ability in schizophrenia

D o schizophrenics smell differently than other people? Indeed, there is evidence to suggest that schizophrenia patients have a unique sense of smell . Despite the fact that schizophrenics have intact olfactory acuity, up to 50% ofmale patients are reported to be impaired on the University of Pennsylvania Smell Identification Test (UPSIT) (1, 2, 3) . This test ofolfactory identification ability includes 40 items, each ofwhich presents a scratchand-sniff patch along with a list of four answer choices . (For example, one item reads, "This odor smells most like : a) chocolate ; b) banana ; c) onion ; d) fruit punch.") Compared to healthy respondents, schizophrenia patients are repeatedly found to demonstrate impaired performance on the UPSIT These findings, however, are based upon between-group comparisons of raw UPSIT scores and raw scores do not satisfy the basic specifications ofmeasurement . Therefore, we decided to analyze the raw score data matrix from a sample of 54 schizophrenics and 133 healthy participants with the Rasch Model for dichotomous observations (4) . The primary goal of our study was to verify that UPSIT items contribute to a single factor with sufficient spread along a discernible line of increasingdifficulty to define a recognizable hierarchy of olfactory challenge . We also examined the clinical utility ofthe UPSIT (i .e., whether UPSIT items separate persons into five distinct levels of olfactory diagnosis as described in the test manual) . Our findings suggest that the UPSIT has succeeded in defining a distinct olfactory identification construct for both schizophrenics and healthy participants . (Item separation indi ces for the SZ and control groups are 1.70 and 2.49, with corresponding reliability estimates of .74 and .86 .)