Quality and reporting of publications by Indonesian researchers: a literature survey.

Objectives: To evaluate the quality of reporting of the risk of bias of the Indonesian medical research.

Methods: Publications from PubMed and non-PubMed indexed Indonesian medical journals between January 2008 to December 2010 were assessed for risk of bias based on criterion combination from Hedges-criteria and the Oxford Center for Evidence-Based Medicine. We assessed whether the publications addressed the risk of bias adequately (quality of reporting) and whether the risk of bias criterion was fulfilled (quality of methods). The quality (both of reporting and of methods) of a study was classified as "high" if, for at least two-thirds of the criteria were adequately reported and fulfilled. It was classified as "low" when only one-third of the criteria were reported and or fulfilled.

Results: Of the 1753 publications, 29% (n = 507) were original medical research. For 21% (109/507) the quality of reporting was high; for 15% (77/507) the quality of methods was high. The proportion of high quality was significantly higher among PubMed than non-PubMed, with difference between proportions: (95%CI of difference: 3 to 23).

Conclusion: A small proportion of Indonesian studies have high quality of reporting or methods. When international reporting guidelines are endorsed and followed, the quality of future studies may improve.