OBJECTIVES: To use item response theory (i.e. Rasch) analysis to develop and evaluate scales to test prescribers' DDI knowledge and perceived usefulness of DDI information sources, and to examine factors that may be associated with prescribers' DDI knowledge. METHODS: Data were obtained from a US national mail survey sent to 12,500 prescribers. The survey instrument included 14 drug-drug pairs that tested prescribers' ability to recognize clinically important DDIs and five 5-point Likert scale-type questions that assessed prescribers' perceived usefulness of DDI information provided by various sources. The knowledge and usefulness questions were examined via Rasch dichotomous and rating scale models, respectively. Regression analysis was used to examine factors related to prescribers' DDI knowledge scores which were derived from Rasch analysis. RESULTS: 950 completed questionnaires were received. Rasch analysis of knowledge and usefulness items revealed satisfactory model-data fit (infit mean square ? 1.5 and outfit mean square ? 2.0) and person reliability of 0.72 and 0.61, respectively. Analysis results suggested that the 14 items targeted well on the prescribers' DDI knowledge levels. Nevertheless, prescribers with little DDI knowledge were targeted by three items that were equally difficult, and these items were too difficult for some prescribers. Among the usefulness questions, the statement regarding the future usefulness was most easily to endorse, whereas the statement regarding how often the information was new to the prescriber was most difficult to endorse. A multiple regression analysis revealed that prescribers' DDI knowledge score was associated with being a specialist and having previously seen a DDI-caused harm. Another significant predictor was the extent to which the risk of DDIs affected prescribers' drug selection. CONCLUSION: IRT analysis indicated that the two scales developed have acceptable reliability and fit statistics. Refinements such as adding easy items to the DDI knowledge test and continuing evaluation are needed in future application of the scales.