Anesthesia and surgery are a risk for all, the healthy as well as the sick. While the prevention of adverse outcomes in healthy patients is paramount, enhancement of safety for critically ill patients is also essential, since they are more likely to suffer a SNO after a critical incident. Dangers originate from a variety of sources, not solely from errors by the anesthesiologist. Simple incidents of all description are inevitable, and we should focus on promoting recovery as well as avoiding error. Processes that lead to negative outcomes after critical incidents should be investigated to reduce the uncertainty complexity associated with managing the human body during anesthesia, and to establish the most effective detection and recovery techniques. Outcome studies are lacking, and clinical and animal research is highly dependent on the chosen model or population, making the results hard to apply to variable clinical conditions. Wherever possible, a consensus should be sought on therapeutic and adverse effects of drugs and techniques in common, specific patient populations. These can serve as a basis for developing therapeutic plans, recognizing that customizing to individuals is always necessary. A mainstay of anesthetic practice already involves attempts to loosen couplings, by keeping homeostatic mechanisms intact when possible (awake intubation, regional anesthesia); providing temporal buffers (titration of drugs, and use of drugs with short onset times and rapid termination of effect); and providing safety margins using appropriate pre-treatments (pre-oxygenation, atropine in children, etc.). Further means of loosening coupling should be identified and promoted. Specific attention to recovery from simple incidents should attack several facets of the problem.(ABSTRACT TRUNCATED AT 250 WORDS)