Introduction
What is new?
- •
Test-treatment randomized controlled trials of diagnostic tests are rare: we estimate that only 37 trials were published per year for 2004–2007.
- •
Randomized evidence of effectiveness of diagnostic tests is unlikely to be available for many tests in specific clinical settings.
- •
Policy and decision makers frequently need to resort to lower grade evidence, such as decision models to provide guidance on test selection and use.
It is recommended that decisions on diagnostic test use should be based on evidence that they do more good than harm in terms of improving patient outcomes. Test-treatment randomized controlled trials (RCTs) are reputed to provide the highest quality evidence for these decisions [1], [2], [3]: patients presenting with a clinical problem are randomized between two or more testing strategies, followed through subsequent diagnosis, management planning, and treatment and evaluated after all interventions have been undertaken.
Guideline developers and systematic reviewers often fail to locate RCTs of diagnostic tests [4], [5], [6], [7], [8], [9], implying that they are rarely conducted [10] or difficult to find, potentially because of the absence or inconsistent use of diagnostic method terms [11]. Search strategies developed to locate diagnostic accuracy studies are known to suffer from the absence of index terms that sort diagnostic evaluations by study type [11]. However, no strategies have yet been designed to locate test-treatment RCTs.
This study aims to estimate the number of test-treatment RCTs published annually.