Background: Reporting bias is a recognized problem in many Cochrane Reviews, and could jeopardize the validity of their conclusions with implications for prescribers and consumers alike. Several currently available Cochrane Reviews of popular antipsychotics for schizophrenia have been criticized due to the inclusion of studies with high dropout rates and limited reporting of results. In an attempt to ameliorate these shortcomings, and to analyze how access to unpublished data may affect systematic reviews, our group is endeavouring to carry out a novel approach to a systematic review of olanzapine versus placebo for schizophrenia by including data extracted from clinical study reports held by drug regulatory bodies. To this end, we contacted the European Medicines Agency to gain access to relevant clinical study reports. We also compared the information in Cochrane Systematic Reviews (CSRs) to summarized information available via the US Food & Drug Administration (FDA) website (drugs@fda).
Objectives and Methods: We submitted an inquiry to the European Medicines Agency, under the access-to-documents policy, seeking access to all clinical reports submitted as part of marketing-authorization applications for olanzapine (Zyprexa) by Eli Lily. We obtained summary reports of the same trials from the FDA website. We compared data extracted to determine differences between each source.
Conclusions: Findings that may be useful to Cochrane authors include the process of requesting access to clinical study reports, limitations on access to data, timeliness of data delivery, and indexing and formatting of data. While the volume of data received is large, and may increase the complexity of the final analysis, access to clinical study reports may improve the reliability of systematic reviews and reduce the effects of reporting bias. Comparison with US FDA review reports also allows us to examine the extent to which full CSRs provide ‘added value’ over summarized information in FDA reviews.