This paper discusses an audit plan, based on a continuous sampling technique, queuing theory and sequential analysis. For a given test operation, if all tested products can successfully pass the operation, only a portion of the arrivals will be tested, that is the process is in an audit mode. However, once a failure occurs, all subsequent products must be tested (i.e. in a screen mode) until K consecutive good products are observed. In this case, the operation will return back to the audit mode. The portion of the products to be tested is determined by the arrival process and the capacity of the test operation. If the test capacity is below the arrival rate, line congestion may occur owing to line screening triggered by a product failure. The degree of congestion is measured by the level of work in process and the actual throughput. Furthermore, products may bypass the test operation under the audit plan. One must carefully estimate the escape rate. Finally, it is always desirable to detect process problems before a large number of faulty products are built. The sequential analysis may be used to establish a real-time problem detection scheme. The same approach can also help to identify test equipment problems. This work was developed for a test operation in an assembly line. It can be implemented either in a computer-controlled environment or in a manual operation mode. Performance of an audit plan is a function of line throughput, escape rate, test capacity and work in process. Numerical examples are presented for tradeoff analysis.