Abstract

One of the most important elements of a quality assurance program is the corrective action process. The success of the corrective action process depends on the commitment of management, but even in organizations where management is totally committed, results may be ineffective or counterproductive. One reason for the ineffectiveness of the programs is failing to consider the effects of variation. W. Edwards Deming demonstrated how management could make a process worse by failing to consider variation in the funnel experiment. Four types of variation categories are commonly used. These are common cause, special cause, intrinsic pattern, and tampering. Tampering is the least-discussed variation in the literature, but it is also the variation that frequently causes well-intentioned corrective action programs to achieve poor results. Tampering involves taking action to correct a situation when only common cause exists, and it is nearly certain to make matters worse. When attempting to avoid tampering, it is important to make data-based decisions and to recognize the difference between correction and corrective action. All problems require correction, but only when it is shown that minor problems of a certain type frequently reoccur is corrective action appropriate. Often, correction action programs in general, and audit programs in particular, force tampering because they require corrective action for minor findings that are not systematic or that represent symptoms and not problems. A better approach would require correction for all deficiencies and corrective action only when deficiencies are found to be major or systemic.