In-band spectrum sensing (SS) requires that secondary users (SUs) periodically interrupt their communication to detect the emergence of the primary users (PUs) in the channel. A new approach referred to as spectrum monitoring (SM) was proposed by Boyd et al., which allows the SU to employ its receiver statistics to detect the emergence of the PU during its own communication periods. In this paper, we construct a decision statistic for SM based on the SU's receiver error count. We then evaluate the detection probability of SM in the presence of interference from the PU signal. Next, we derive closed-form formulas for channel utilization and detection delay using two Markov chain models. Upper and lower bounds on channel utilization and detection delay are derived, and an optimization problem is formulated and solved to maximize channel utilization with a constraint on detection delay. Numerical results from analysis are compared with simulation results obtained for a BCH code and a convolutional code, which show the accuracy of the analysis and the significant improvement of a hybrid SM/SS over SS alone.