BACKGROUND: The literature retrievals are still the main problems which limit the quality of systematic review (SR) for traditional Chinese medicine (TCM), and there were still no comprehensive analyses for these problems. The aim of this study is to analyze these problems to guide future SRs for TCM. METHODS: The Cochrane Library (Issue 2, 2009) and the Chinese BioMedical Database (CBM, 1978- 2009.7.31) were comprehensively searched to identify the SRs of TCM. The native literature retrieval rate of the first author, the characteristics of the included databases, the number of the included database, the reporting of literature retrieval strategy, and the reporting of adjuvant search for both the SRs from the CBM and Cochrane Library were analyzed. RESULTS: A total of 341 SRs were included in our analyses, including 245 from the CBM and 96 from the CL. The Chinese-authored SRs ranked the first in median retrieval rate (100%). The Cochrane SRs were significantly superior than the CBM SRs: more searched databases (media: 4 vs. 6, P < 0.001), more precision of retrieval time, more conducting gray literatures searches (47.92% vs. 20.41%) and handing searches (57.29% vs. 54.69%), and more reported retrieval strategies (66.67% vs. 8.16%). CONCLUSIONS: There were significant problems in the literature retrieval of SRs for TCM. In future, when performing any SRs for TCM, sufficient and proper databases, clearly searching strategies and times, proper adjuvant retrievals should be reported. Cochrane SRs might be a better choice.
BACKGROUND: The literature retrievals are still the main problems which limit the quality of systematic review (SR) for traditional Chinese medicine (TCM), and there were still no comprehensive analyses for these problems. The aim of this study is to analyze these problems to guide future SRs for TCM. METHODS: The Cochrane Library (Issue 2, 2009) and the Chinese BioMedical Database (CBM, 1978- 2009.7.31) were comprehensively searched to identify the SRs of TCM. The native literature retrieval rate of the first author, the characteristics of the included databases, the number of the included database, the reporting of literature retrieval strategy, and the reporting of adjuvant search for both the SRs from the CBM and Cochrane Library were analyzed. RESULTS: A total of 341 SRs were included in our analyses, including 245 from the CBM and 96 from the CL. The Chinese-authored SRs ranked the first in median retrieval rate (100%). The Cochrane SRs were significantly superior than the CBM SRs: more searched databases (media: 4 vs. 6, P < 0.001), more precision of retrieval time, more conducting gray literatures searches (47.92% vs. 20.41%) and handing searches (57.29% vs. 54.69%), and more reported retrieval strategies (66.67% vs. 8.16%). CONCLUSIONS: There were significant problems in the literature retrieval of SRs for TCM. In future, when performing any SRs for TCM, sufficient and proper databases, clearly searching strategies and times, proper adjuvant retrievals should be reported. Cochrane SRs might be a better choice.