Awareness of the need to continue the safety assessment of new drugs into the post-marketing period has continued to increase over the past few years. The stage has been reached where there is, for all practical purposes, unanimous agreement that all new drugs, particularly those intended for widespread use and/or long-term administration, be subjected to some form of post-marketing monitoring. In many cases nationally sponsored spontaneous reporting systems are claimed to be sufficient to meet the needs without any critical appraisal being made of the suitability or reliability of this method. The adoption of 'pharmacovigilance' (spontaneous reporting) as the system for the whole European Community (EC) emphasizes the faith that has been placed in the system. At a recent symposium (Management Forum, 9-10 March, 1992) on post-marketing safety evaluation, more than half of the 16 papers presented were entirely or predominantly concerned with spontaneous reporting systems, which is a pattern common to other meetings on this topic. The Committee on Safety of Medicines (CSM) has referred to spontaneous reporting as the cornerstone of post-marketing surveillance which has further endorsed a general feeling of confidence that the system will meet our societal needs even though its scientific capability may be deficient. It is the purpose of this paper to challenge these assumptions and to make an objective appraisal of the various methods available.