Challenges of identifying unpublished data from clinical trials: Getting the best out of clinical trials registers and other novel sources

Abstract

Clinical trial data are essential for assessments of the effectiveness of health care interventions. Information about ongoing or completed, but not yet formally published, trials has been more difficult to identify until the development of clinical trials registers and portals. This paper summarises research evidence on identifying sources of trial data, how and when to search those sources, and which future developments may enhance access to and retrieval of unpublished trial evidence. We conducted a literature search for relevant studies and provide a narrative review of the evidence from these studies. Clinical trial data can be found in resources including clinical trials registers, regulatory agency sources, health technology assessment websites and manufacturers' websites, and submissions for regulatory approval. The challenges of searching these resources are described. Trials registers are relatively unsophisticated in terms of their search interfaces, and searchers need to adapt to each individual register. There is overlap across registers, but little research on the degree and nature of overlap and how best to search. Despite these challenges, trials registers and other resources can be rich sources of additional unique trial data, which may not be available from journal reports. New initiatives, such as OpenTrials, aim to consolidate and link all structured data and documentation related to clinical trials. No single resource gives access to all trials, and multiple registers should be searched as sensitively as possible. Searching is challenging and should be adequately resourced. Information specialists should monitor new developments which may reduce the challenges over the coming years.

State of the art, state of the science?

“…Collaborations span institutions and the academic‐commercial divide . Academic review organizations (eg, ScHARR, PenTag, and YHEC) and health technology agencies (eg, CADTH) and networks (such as Cochrane , and the Medical Library Association collaboration) continue to play a critical part. Interest in sources and in indexing for retrieval persists through the decades.…”

“…Interest in sources and in indexing for retrieval persists through the decades. Although, understandably, a preoccupation with trial reports continues to drive the agenda, presaged by the 1994 BMJ article, this Special Issue reflects wider interests in trial protocols, unpublished data, and qualitative studies . The latter addition is particularly welcomed by me, as a Cochrane Qualitative and Implementation Methods Group coconvenor, given that qualitative research remains a poor relation in terms of research outputs and research funding.…”

Graphical augmentations to sample‐size‐based funnel plot in meta‐analysis

“…Although using unpublished data can greatly remedy publication bias, it may not be always feasible for all meta‐analyses. Also, it often takes considerable time and resources to obtain unpublished data, and the quality of such data may not be as good as the published data in peer‐reviewed journals. Therefore, statistical methods for publication bias have been widely used in meta‐analyses …”

Innovation in information retrieval methods for evidence synthesis studies

“…The expansion in the definition of evidence in terms of both format and research tradition is addressed by three contributions. Sutton et al and Isojärvi et al consider the challenges of retrieving unpublished evidence, where searching cannot rely on the standardized bibliographic format of meta‐data and abstracts, or the organized navigation offered by the controlled indexing of database thesauri . Isojärvi and colleagues review in detail sources of unpublished trial data and consider the issues associated with how these important sources might best be interrogated.…”

“…Sutton et al and Isojärvi et al consider the challenges of retrieving unpublished evidence, where searching cannot rely on the standardized bibliographic format of meta-data and abstracts, or the organized navigation offered by the controlled indexing of database thesauri. 20,21 Isojärvi and colleagues review in detail sources of unpublished trial data and consider the issues associated with how these important sources might best be interrogated. The paper by Sutton et al focuses on clinical trial protocols, comparing and evaluating different search approaches for identifying this specific form of unpublished evidence.…”