As the basis of conducting real-word studies

the validity of real-world data（RWD） should be highly valued. RWD was characterized by the universality of sources

the abundant of data

the richness of content

and the diversity of types

and was highly dispersed and heterogenized. The risk factors that might lead to bias of RWD were analyzed from the aspects of data source and data statistical analysis. To improve the RWD validity

the quality of medical records should be improved; patients′ education and researchers′ training are needed; the system should be perfected and the data analysis should be standardized.