Background: The ultimate aim of diagnostic test evaluation is to determine which tests have the most favourable impact on patient health. Randomised controlled trials are the methodological gold-standard for evaluating these questions, however it is a common anecdote that such publications are rare . But how rare are they? To date none has attempted to enumerate the true extent of this important evidence-base. Objectives: To estimate the number of test-treat randomised trials published within a defined time-period, as indexed in the Cochrane Central Register of Controlled Trials (CENTRAL). Methods: A capture-mark-recapture technique was used to estimate the total number of test-treat trials published between 2004 and 2007. CENTRAL (Issue 2, 2010) was searched using two separate ascertainment strategies. The first (S1) used terms related to diagnostic research, and was not restricted by test modality. The second (S2) was a test-specific search that targeted the names of five mainstream imaging modalities. Studies were included if they randomised patients with suspicious signs/symptoms to receive a diagnostic test followed by the administration of treatment. Trials that did not assess patient outcomes were excluded, as were those that evaluated serial testing or population-based screening. Each search was screened independently by two researchers. Results: Search strategies yielded 13,576 (S1) and 12,041 (S2) titles and are currently under review by the second screener. Upon completion, findings will include a 2x2 contingency cross-tabulation of the number of relevant studies retrieved by both or only one of the searches from which an analysis of ascertainment overlap will produce the estimate of total number published. Agreement between screeners will be presented as kappa values. Conclusions: The estimate as a proportion of all trials indexed on CENTRAL will be discussed, and challenges encountered in constructing effective strategies for identifying test-treat trials commented on.