Testing a tool for the classification of study designs in systematic reviews of interventions and exposures showed moderate reliability and low accuracy.