Sensitivity and reliability are two important issues for membrane integrity testing. Sensitivity of a membrane direct integrity-testing method is likely to govern the removal credit granted to a membrane system. The recent data from a microbial challenge test on production plant scale validated the sensitivity of integrity-testing method the first time and demonstrated high sensitivity of the pressure-hold test for a membrane system. Reliability, defined as the ability of an integrity-testing method, has not been a regulatory focus. However, it is a crucial issue when establishing upper control limit (UCL) for indirect integrity monitoring. The reliability of a membrane integrity-testing method can be evaluated via probabilities of false negative and false positive signals generated by a method and monotonic response to increasing severity of membrane integrity-breach. Because of randomness of a measurement, an UCL has to be established based on some statistical characteristics of a sample population. Recent data indicated that many of the indirect integrity-testing methods are not only low in sensitivity, but also unreliable, which makes indirect-integrity monitoring rather difficult to implement. Includes 9 references, tables, figures.