Objective: To improve clinical utility of the Nurses Observation Sca le for Inpatient Evaluation (NOSIE-30) in China. Methods: Based on the original scale, which only provides names and ranked scores for items, we developed clear definitions for each item and anchoring descriptions for score rating. 164 firs t-episode schizophrenic patients were evaluated with this scale every two weeks after admission. A total of 994 evaluations were completed. The collected data w ere randomly divided into two independent data sets. The number and content of t he subscales were revised based on principal component factor analysis and item analysis of the first data set. The revised scale was then evaluated and compare d with the original 7-subscale version of the instrument using the second data s et. Results: Factor analysis found that the factor structu re of the revised scal e was markedly different from that of the original instrument. Item analysis res ulted in elimination of 30 original items and redistribution of the remaining 26 items into 5 subscales. Comparison of the revised scale with the original instr ument revealed better psychometric properties for the former than the latter.Conclusion: The subscale scores of the original 7-subscale v ersion of NOSIE-30 did not reflect independent symptom clusters in Chinese schizophrenic inpatients. T o improve the usefulness of the scale in China, the present study recommended a better revised scale with re-definition of items into 5 subscales.