Background: Meta-analyses of controlled clinical trials that summarize existing evidence by statistical methods make up an integral part of a systematic review and provide systematic background knowledge. The extent to which investigators use existing evidence efficiently in the design of clinical trials and report their results in context has been questioned. The recently developed sequential meta-analysis methodologies are attractive because they provide longitudinal trend of effects, control for random errors and multiplicity correction. Sequential Meta-Analysis Research Tracks (SMART) is proposed to incorporate the sequential methods into the current meta-analysis procedure.

Methods: We integrated basic concepts and methodologies of meta-analysis in this poster. SMART was represented as a flowchart in Figure 1.

Results: Two working examples were selected to demonstrate how SMART could be used to glean information for future studies and gain earlier awareness of efficacy and adverse effects. In the selenium example, SMART could be used to glean information for future studies thus circumventing ineffective trials and improving the benefits of patients. In this case, SMART showed that future trials with selenium should have long treatment duration. In rosiglitazone example, once the sample sizes had reached the total information size, there will be no need to carry out further trials, which might bring harms to patients. After conclusive results had been obtained, further effort should be focused on either reducing the harms of patients or increasing the benefits of patients.

Conclusions: The applications of SMART demonstrate that we can glean useful information for future studies thus circumventing ineffective trials and improving the benefits of patients. With the increasing importance of meta-analyses in evidence-based medicine, we should maximize the use of this powerful tool to greatly improve human health.