Skip to main content


Comparison between Autocad and Draftsight in data extraction

Date and Location




Sunday 22 September 2013 - 10:30 - 12:00


Presenting author and contact person

Presenting author

Zheng Li

Contact person

Deng-Hai MI
Kehu Yang
Abstract text
Background: None included studies in Systematic reviews can be lost to perform meta-analysis. We can use autocad and draftsight software to gain the data we needed when published figures as the only source of outcome data. Draftsight is more convenient for using to extract data than autocad. However, is it accurate and reliable enough performing draftsight for data extraction? Objectives: To analyze and compare the precision of autocad and draftsight software for data extraction from figures by performing meta-analysis. Methods: We prepared ten figures together with original data from published papers, which would be used for the precision test. Mean, 95% confidence interval (CI) and standard deviations (SD) were used as statistical parameters from figures to elevate the data extraction. Autocad and draftsight were performed by different researchers, and data analysis was implemented by the third researchers, which were designed with blind method to avoid bias. Differences in data extraction between researchers with the same software were previously compared and adjusted. Meta-analyses were performed to compare precision between the two software, using original data from papers and the data extracted with autocad and draftsight. Results: We gained the data exactly from figures using the software. As result of the meta-analysis, there was no statistically significant difference of precision between autocad and draftsight software for data extraction; and both of them showed their satisfactory precision. Conclusions: There is no difference of accuracy between the two software; however, draftsight is small and exquisite in the practicality, which is more convenient than autocad. So we would prefer to recommend draftsight in general. With the draftsight software, we can always get the data exactly when papers were published with figures as the only source of outcome data. Then we can perform meta-analysis without data lost of included studies.