Successful spectrum sharing in a cognitive radio network depends on the correct and quick detection of primary activity. Cooperative spectrum sensing is therefore suggested to enhance the reliability of such detection. However, it renders another significant problem of increased detection delay and traffic burden. Moreover, efficient schemes for multi-sensor data fusion should be designed. In this paper, we employ sequential detection scheme together with ordered transmissions from cognitive detectors. We derive two sequential schemes, one that is capable of achieving the minimum probability of error, and another that trades-off performance with delay. For the latter we derive expressions for the likelihood functions of ordered observations and compute the thresholds via backward induction. Simulation results demonstrate the relative performance of the approaches proposed in the paper.