Objective: To investigate epidemiology characteristics, methodological quality and statistical reporting of network meta-analyses (NMAs) in the field of cancer.
Methods: Twelve databases were searched from inception to 9 July 2015, to identify any NMAs (including adjusted indirect comparison) in the field of cancer in English or Chinese languages. We assessed general characteristics, reporting of the literature search, reporting and quality of statistical analysis, and assessed the methodological quality using a modified AMSTAR (A Measurement Tool to Assess Systematic Reviews) checklist. Reporting quality of statistical analysis and methodological quality was stratified by general characteristics, and analysed by Chi-square test using STATA version 12.0.
Results: From 6408 citations retrieved, we identified and included 102 NMAs in the field of cancer, including 92 NMAs published in English and 10 in Chinese. Forty-three per cent of the included NMAs had been published since 2014; 98 NMAs involved 24 different cancers, and four NMAs did not specify the types of cancer. Non-small cell lung cancer was the most common cancer to be studied in the included NMAs (19%). NMAs were most often performed by researchers based in China (28%). The median publishing period was 101 days (inter-quartile range (IQR): 47 to 187 days). The median total AMSTAR-score was 8.00 (IQR: 6.00 to 8.25). Methodological quality and statistical reporting did not differ substantially by selected general characteristics.
Conclusions: The methodological quality of NMAs in the field of cancer was acceptable. However, some methodological flaws have been identified in the published NMAs, especially regarding searching of literature, assessment of scientific quality, appropriate consideration of scientific quality in formulation of conclusions, the methods used to synthesize findings of studies, and assessment of publication bias.