Article type
Year
Abstract
Background: Previous recommendations about information to include in clinical trials results databases would require a substantial investment in human labor to collect and maintain across all trials. A smaller list of very useful elements would be a practical way initially to construct a clinical trials results database with broad coverage.
Objectives: To develop a list of items that systematic reviewers consider the most important elements to include in a clinical trials database; to use these as a guide for recommendations about the design of a clinical trials results database.
Methods: We invited experienced systematic reviewers to participate in a discussion group. Experts applied their experiences to the question of what information would be most essential in a clinical trials results database. We then asked experts to rank these database elements in terms of usefulness in their daily work.
Results: Experts viewed the clinical trials results database as a supplement to currently accessible information (e.g. journal publications). The following elements had wide consensus among the experts as being valuable in providing additional support in conducting comprehensive systematic reviews:
- Drug name;
- Trial name;
- Disease or condition investigated;
- List of all trials that have been conducted per drug;
- Identification of unpublished trials;
- Intervention;
- Primary outcomes;
- Number analyzed for primary outcome;
- List of all publications per trial;
- Unique trial identification number;
- Number randomized;
- Results of primary outcomes.
These comprise the set of elements that are likely to be useful to the greatest number of systematic reviewers. Additionally, important usability issues were uncovered, particularly the ability to do full searching on database fields and perform free-text searches within linked documents.
Conclusions: Focusing efforts on these elements and features would allow a balance between broadly useful results information, manageable data collection, and high functionality.
Objectives: To develop a list of items that systematic reviewers consider the most important elements to include in a clinical trials database; to use these as a guide for recommendations about the design of a clinical trials results database.
Methods: We invited experienced systematic reviewers to participate in a discussion group. Experts applied their experiences to the question of what information would be most essential in a clinical trials results database. We then asked experts to rank these database elements in terms of usefulness in their daily work.
Results: Experts viewed the clinical trials results database as a supplement to currently accessible information (e.g. journal publications). The following elements had wide consensus among the experts as being valuable in providing additional support in conducting comprehensive systematic reviews:
- Drug name;
- Trial name;
- Disease or condition investigated;
- List of all trials that have been conducted per drug;
- Identification of unpublished trials;
- Intervention;
- Primary outcomes;
- Number analyzed for primary outcome;
- List of all publications per trial;
- Unique trial identification number;
- Number randomized;
- Results of primary outcomes.
These comprise the set of elements that are likely to be useful to the greatest number of systematic reviewers. Additionally, important usability issues were uncovered, particularly the ability to do full searching on database fields and perform free-text searches within linked documents.
Conclusions: Focusing efforts on these elements and features would allow a balance between broadly useful results information, manageable data collection, and high functionality.