How health technology assessment agencies address the issue of unpublished data

Reporting bias potentially threatens the validity of results in health technology assessment (HTA) reports. Our study aimed to explore policies and practices of HTA agencies regarding strategies to include previously unpublished data in their assessments, focusing on requests to industry for unpublished data.

We included international HTA agencies with publicly available methods papers as well as HTA reports. From the methods papers and recent reports we extracted information on requests to industry and on searches in trial registries, regulatory authority Web sites and for conference abstracts.

Eighteen HTA agencies and seventy-three reports were included. Agencies' methods papers showed variability regarding requests to industry (requests are routinely carried out in seven cases, not mentioned in six, at the discretion of HTA authors in three, and based on manufacturer applications in two), which were reflected in the reports investigated. As reporting of requests was limited, it often remained unclear whether unpublished data had been received. Searches in trial registries, at regulatory authorities or for conference abstracts are described as a routine or optional part of the search strategy in the methods papers of 9, 11, and 8 included agencies, respectively. A total of 52 percent, 39 percent, and 16 percent of reports described searches in trial registries, at regulatory agencies, and hand searching of conference proceedings.

International HTA agencies currently differ considerably in their efforts to address the issue of unpublished data. Requests to industry may constitute one strategy to access and include unpublished data, while agencies can learn from each other concerning successful practice.